Start with a Website, Not a Mobile App | Hacker News Reader

Start with a Website, Not a Mobile App

atrium.co

1,620 pointsjenthoven559 comments

80% of apps are uninstalled in the first 10min (especially if they ask for information or creating an account). Once uninstalled very unlikely they will ever install again.

In my dev shop, I always advise clients to look into website MVP over apps unless IT IS CRITICAL to the functionality. Websites can always bring back ppl who were initially uninterested. Faster development/instant updates. Better tracking and UX anazyling.

Edit: also better collaboration tools.

I completely agree. This is also my advice to clients that initially ask for a mobile app as well.

Unless you _know_ you have a successful business, a website/webapp MVP is almost always a cheaper way to validate your business concept and a cheaper way to iterate as you develop it.

How many websites ever get passed the signup stage? I often will leave as soon as I'm tasked with creating an account and I'm not likely to come back.

What is especially frustrating that many sites don’t allow Mailinator and other similar ‘one-time’ email systems to be used for registering an account. No way I am going to give you my email address up front before seeing the product.

Now the solution is of course to have a separate gmail account to register. Most sites still allow user+site@gmail.com address syntax so that I can sometimes see who are leaking where.

I would be curious to know if there is some credible A/B testing showing that for a relatively unknown new product blocking Mailinators and it’s kin somehow results as a better conversion?

There is a lot of temporary email services constantly popping into the Internet every month. Some of them are open-sourced so one can make own quick disposable email or forwarding service with cheap domain. "Resistance Is Futile" - no way any exclusion list can keep up to cover them.

every time i do a search i find services that require registration or even payment, and rarely ever anything that supports making up an email address on the spot the way mailinator does.

Given how well known user+site@gmail.com is wouldn't many sites simply strip the +... part before selling the email data?

Yes, this is part of standard data cleaning practices. You always remove whatever custom formats a provider allows to get the raw minimum email address (this is only for export, you still send official communication with a user to their original specified address).

There's a simple solution for that: always include a label as part of your address, and treat any mail that doesn't include one as spam.

Not every service even allows you to have a label in your email address. So what then?

Get them to fix their broken systems? The concept of a "label" exists only in the user agent. To everyone else the '+' character is just part of the account identifier, and a perfectly valid element of any e-mail address. The entire portion before the '@' sign is potentially significant. If someone hands you an address abc+def@example.com, there is no guarantee whatsoever that mail sent to abc@example.com will end up in the same individual's inbox—that could be a completely unrelated account.

On the e-mail provider's side, one way to work around such broken systems would be to let users generate random, opaque aliases which do not contain their primary e-mail username but route to the same account, with a predetermined label applied on receipt. These would look just like ordinary addresses and there would be no way for services to strip out the identifying elements.

Why would they care enough to do so?

If I get a marked as spam notification, I'll strip anything after the '+' (and really any special characters that distract from the raw email address) to make sure I never send an email to the address again.

There is one surprisingly good feature in yahoo mail that I wish google would provide: the ability (requirement, actually) to create a custom prefix for disposable email addresses. Just pick something completely unrelated to your real username. You can only create it once and it is forever your prefix, so don't mess up this step! Also, unlike gmail, you have to register a suffix when you want a new address in the form of prefix-suffix@yahoo.com. But this has the added benefit that you can delete any of the suffixes from the list, thereby stopping incoming messages from that address. You can have up to 500 active suffixes.

Finally, since the format uses a dash rather than +, I have never once had a disposable yahoo address rejected.

Not sure if that made any sense, but here's a link to the docs:

https://help.yahoo.com/kb/SLN28338.html

Tangent: I've been using a nispam@inboxalias.com for years, and it's only recently that I've encountered validation rejections. (Ironically, both times that happened, the form in question did allow +foo@gmail.com.)

Yup. It's frustrating.

One benefit of switching to mail on my own domain (powered by FastMail at the moment) is that instead of username@mydomain, I can also use antispamlabel@username.mydomain. Which is hopefully less common and too dangerous to automatically filter out in "data cleaning".

Man I so want to do that, but the pricing just seems outrageous at minimum $3/user/month, standard $5/user/month.

I've been wanting to hack together some scripts to use a serverless cloud provider like SES. Since I rarely check my email and almost never send any, it's crazy to pay such a high monthly fee.

I use mailgun to forward mails incoming on my domain to my usual web mail accounts. Free tier is more than enough, since i don't expect hundreds of mails per day.

Sending is a bit painful, though. Also sometimes mails fail to be forwarded and get stuck for a while... annoying when you're waiting for some kind of link / code.

Interesting, thanks for the tip!

If you get a lot of spam and it's all forwarded than some providers like Gmail will treat your email address as the source of the spam rather than the original sender.

Register a domain name with mail forwarding and use wildcard aliases to forward all mail on your domain to your real mailbox.

Your email is public, there's no gain in keeping it secret.

For further reading: https://www.troyhunt.com/im-sorry-but-your-email-address-is-...

Yes but that doesn't mean I want my inbox receiving emails a year later from a service I used for 10 minutes. I generally trial services (that don't contain any personal information) with a separate mailinator email. If I feel like this is something I'm going to be using more of, I'll do a proper registration with my actual name and email.

> I would be curious to know if there is some credible A/B testing showing that for a relatively unknown new product blocking Mailinators and it’s kin somehow results as a better conversion?

Well, there's always this: https://articles.uie.com/three_hund_million_button/ -- trigger warning: that page has annoying delayed-by-few-seconds modal popup that begs you to subscribe and, incidentally, hand out your email address; the irony must have been lost on their marketing folks.

Unless your business is REQUIRED[×] to have user details on file, putting any kind of friction on the user's action path is equal to voluntarily giving up sales. No, I don't want yet another account on your shitty website, no matter what it sells to me. If I end up coming back often enough, then I may, voluntarily, create an account. But if your site requires to create one up front, I'm incentivised to take my business somewhere else.

The worst offenders try to do email domain TLD matching against country of residence. I've seen it only twice, but it's such a user-hostile pattern that it has seared itself to my memory (while mercifully having relieved me of the memory of where I ran into it).

×: cough gambling, financial services, insurance, etc.

What is it that makes you actually sign up for a service?

Either I'm already certain or it's necessity/desperation.

If I'm actively looking for an certain type of application I will usually dedicate half a day to sign up for bunch of different apps all at once. It's pretty time consuming and I'm usually left disappointed. I will then spend the next half a day trying to remove myself from them all.

Then 6 months later one of those app companies send you an email update, by which point you've forgotten what they do, and yet they give you no hint of what the app is in said email.

Or they email you 3x a day for something you were interested in for no more than 10 minutes.

Or year later they got bought by some other company and your account is transitioned to the new company but those emails automatically got sent to spam and now all of a sudden you’re a customer of a new company and you had no idea. It’s happened to me more than once.

Thank you! I personally think that mobile web and app design has been one of the greatest atrocities in tech. Not only is it trash on mobile, it's spread and made lots of sites trash on desktop as well because people seek a one-size-fits-none solution.

I personally go by the rule that if your service can work as a website, it should be a website. Not ever a standalone app unless there are important requirements which can't be met in a browser. That can only be good for feature creep, gimmicks, and of course data scraping and other violations. A lot of companies end up purposely hamstringing their website--or push you into a terrible mobile site and ignoring your browser's request for the desktop version, plus cover half the screen with multiple instances of begware. Yes I'm looking at you reddit, you filthy old turd. It's incredibly user-hostile. I'm sure it drives plenty of cheap, chintzy short-term engagement but y'all are going to perish like digg sooner or later.

I wish companies would start really putting forward desktop deployment as a major priority again. Maybe I'm old and behind the times. But it seems like all serious activity in my life is always done on the desktop. As soon as I encounter something where I'm going to be doing a lot of browsing or typing or any effort, I move to the desktop platform where the interface and experience are just plain better in every way. I hate interacting through a terrible, tiny touchscreen by swiping around like a finger-painting toddler or using a minuscule keyboard which autocorrects me into apparent illiteracy. If I'm not out in the field and have no other choice, I will always, always choose the desktop environment for its many enormous advantages over mobile. Mobile is here to stay, but the UX still sucks enormous ass and is only tolerable because it's ubiquitous, plus everybody sucks relatively the same from mobile platform to mobile platform so it kind of seems normal.

Once again, maybe I'm just old and now I'm like my mom who stubbornly sticks to VIM in CLI and resents anything that requires clicking a mouse. On the other hand, I kind of see her point and totally respect her cantankerous old-fashioned style of computer literacy. She may hate modern interfaces, but she still knows how to do plenty of fundamental things that even young people these days are clueless about. I believe that simplistic, overly-intuitive interfaces can cause people to live up to expectations and become worse with actual computer literacy, that the golden age of power users is fading, but that's another story.

I would love to have a phone that runs a desktop version of linux and has a slide-out keyboard like they had in the 2000s or whenever that was. Something that just sticks to a desktop user agent and runs desktop executables. Perhaps even something that runs x86 and has a swappable battery to compensate for its power inefficiency. Basically a desktop computer disguised as a mobile device that fits in your pocket. Not iOS or Android or some rooted variant of Android. I don't even give a damn if it has voice calling or not. I think that would have enormous potential and would command a small yet incredibly dedicated userbase, but so far I haven't found anything quite like it.

I love that your mom is a VIM users. That caught me off guard and I lol'd. My mom: not a vim user (or emacs for that matter)

Technically she uses original VI on a Sun Solaris. That thing is still kicking. It's delightful, really.

Id suggest a cloud VM except for bandwidth/latency issues.

You might really like this device, it runs a real desktop Linux and has a really nice keyboard. Its still ARM but meets everything else: https://store.planetcom.co.uk/products/gemini-pda-1

That looks a really nice and very usable modern incarnation of the amazing Psion.

Where does this metric come from?

Many folks that i know of don't keep browsing App stores for apps (some may look for new games now and then) and i wonder it may apply to substantial number of users out there. The app has to be some how popular (either by word of mouth or ) before it gets installed, where as browsing web sites is just a matter of clicking some link.

Just my observation but i rarely see or know of mobile users who browse app stores without first being taken then by a wish/link to download an app. Users who independently dive into app stores are mostly developer-oriented people looking to solve a problem with an app.

Apps go viral via memes or social channels and most of that is thus other apps.

Web, other than being much simpler, and easier to find, has the important benefit of not requiring a user to change context - they can switch tab but they are still in a browser - so their personal 'flow' is far less interrupted than when they change to another app. That interruption has flow on effect to retention, productivity, integration.

Bottom line is for a startup, esp a bootstrapping one, start with web MVP, get it out there fast even cobbled from existing apps and services. Prove your hypothesis as quickly as possible - and when data proves otherwise go pivot. Rinse-repeat.

Where did you get that 80% from? I have hard time believing the average Joe uninstalls an app

The figures are in the order of magnitude that OP suggests, but only for a) games that b) are ad supported (which is how we know the numbers).

More serious apps get to stay a few extra days from the looks of this blog post (among other similar promotional ones on the topic that are a few googles away):

https://www.adjust.com/blog/unmasking-uninstalls/

Seeing how e-commerce, travel, and health are the only three categories that get more than a week I'd spitball it's because users of these apps kept them for a one-time event.

Either way, OP is correct in that as an app publisher, you're only one shot at getting that precious first impression, and odds are your new users will only keep the app around long enough to open it a few times; as few as once, in the case of games. Which sounds about right if my own anecdotal usage is anything to go by.

As an illustration, I just downloaded and uninstalled the waiter.com app. Screw that, I’ll use the website.

This would work unless, the client wants to monetize via freemium model; in that case an IAP with native app would be far more intuitive than a web app as integrating a payment gateway which accepts global payments (Stripe is not available in all countries) is harder than integrating In app purchase on iOS or android.

Where does that 80% stat come from? Do you have hard data or just gut feeling?

> Faster development/instant updates. Better tracking and UX anazyling.

That’s a horrible argument in my opinion. Yes, it is more convenient from a developer’s or a business point of view. But what if the user sees more value in an app than a website, shouldn’t that be the first thing to consider?

I think it's more of an added benefit to building a website first. Especially when trying to validate new ideas which is where the better tracking / UX analysis comes into play.

The web is the fax machine but it’s cool you advise your clients to use a dying technology.

What about the market caps of the top three tech companies combined makes you think web is the future? Just curious.

Could you point to research related to the 80% or similar stats?

I am in an internal conversation regarding building an app vs a website and would love to have some ammo for my arguments.

Consultant is trying to sell an app and to me that looks crazy given that a website could fulfill the same purpose without platform fragmentation, installation etc.

Not OP but I use multiple sources when doing consulting on mobile strategy. Not sure if every source allows to copy/paste data here so I'll just name the sources here :

- comScore MMX Multi-Platform, January 2017

- Deloitte Global mobile Consumer Survey, May 2016

- eMarketer, App Marketing 2015: Fighting for Downloads and Attention in a Crowded Market, July 2015

- Selio user acquisition costs

I want to agree with you, but how many people are regularly surfing the web?

I actually think most people still are, but I keep getting this impression from HN and others in the tech industry that "everyone is on their phone and using apps", and I've yet to see any convincing research into the use of web browsers over the years. I don't know what I'm really supposed to think, but it seems that a lot of us in the industry are convinced that mobile browsing is "dead" or at least sits on the bleachers while apps do basically everything and use the browser as a webview for minor things.

I'd love to believe that most people aren't merely sitting on Instagram and Reddit all day, never touching their browser.

For the life of me I cannot imagine how painful it must be to spend your time only in apps like Reddit and Instagram. That sounds painfully boring and isolating.

I am the exact opposite as I spend all of my mobile time in the browser.

It's a complex study to perform it with any kind of accuracy. And if there's any concrete studies done this already then I would love to see citations.

How do you start your browsing day? I always start from the twitter or reddit app. Most people don't even know the "web", they just stay on Whatsapp, Instagram and Facebook the whole day. (I wish this wasn't the case since this makes so difficult to little people to compete on the space).

Not the OP but I strongly favor browser over apps whenever possible. I’m a news junkie so my progression usually goes:

Washington Post (I live in DC), WSJ, NYT, Hacker News, Maybe Facebook, If really bored, will compare Drudge Report / HuffPo / Fox News just to see the extremes,

Back to HN, eventually feel browsing guilt and go do something productive

I you read Reddit most articles contain a link to some website. That is what Reddit (and HN) discussions are about. So you still need to browse the web even if you start in an app.

I don't think it matters necessarily. It depends on the category your product falls into, and you have to look at how similar products acquire their users.

What software have you purchased in the past year? For my pesonal sample size of one, most of my non-game purchases were business productivity SaaS. And the decision generally happened after repeated exposure via word of mouth and/or discussion forums.

I definitely didn't discover any business productivity software in an app store. And signup that requires any significant amount of information is harder on mobile. I bet SEM and getting customers through Google still works online if you can pull it off.

I've got many clients who aren't tech savvy but come to me talking about "mobile first". But after a brief discussion they've all agreed to go website first and it has saved them a lot of money.

Regarding uninstalled apps, during a productivity self-study last year I deleted probably 75% of the apps off my phone, and cleaned the home screen to only include 7 apps I use daily. Highly recommend it if you want to reduce distractions.

I don't even bother installing apps anymore, I don't have time to research whether or not they are going to abuse my privacy or have some horrendous TOS. I got my core 10 apps and haven't bought a new one in a year.

This is my position as well. Unless your app is going to provide substantial features over your website or operating system like functionality. I'm not even going to try it out.

I don't go to new websites anymore either. Just like I have a handful of core apps, I have about he same number of websites I visit. I think the only webapps I use are Workflowy (love it) and Slack (hate it).

Workflowy has updated their app for Android and it's really good!

I'm a pretty heavy user of it, and it was a godsend to have a nice official Android client that works well. Make use of the web lots on the desktop but it's probably 50/50 desktop/mobile. Love them both .

The only two things workflowy is missing is (1) being able to embed pictures and (2) being able to hyperlink to other nodes to make it act like a graph. It'd be the absolute perfect tool if it had that.

I only use it on the desktop and there I use it in the browser. They have a desktop app but it isn't very good.

Embedding pictures sounds like a nice idea, but I can't imagine how they could do it and have it look good. If I were them, I probably wouldn't add images.

I kind of hope they resist the urge to keep adding features to Workflowy. It feels done to me and I like simple, focused tools.

I like this concept. Using the smart phone more like a PDA than a tiny laptop.

What are your 10 core apps if you don't mind sharing?

For me:

Personal Capital (finance), WhatsApp, Slack, Overcast (podcasts), Kindle, Dark Sky (hyper accurate weather app), Google Authenticator, Lyft, RENPHO (my smart scale’s app), 1Password

Everything else is very situational and totally optional.

off topic of the thread: do you have a strong opinion of Dark Sky vs Weather Underground?

I have found Dark Sky is great in populated areas (NYC, Denver, Houston), but Wunderground seems to do better outside of the metros (Lower Hudson River Valley, Catskills, Adirondacks -- you can probably tell where I live and work, now).

Dark Sky : T-Mobile :: Wunderground : Verizon

Got it.

I've found the best data comes from the nearest/best doppler radar(s). Often this is one of the local tv broadcast stations. Obviously this radar data is licensed to various weather service providers but are usually owned by a broadcast television network operator.

I really like Weather Undergrounds local stations. Being able to see that a location at the top of a hill in town is a couple degrees cooler or windier than one down the street has proven to be really useful when decided to bring a jacket or not. That said Dark Sky's UI is top notch and the data seems to be pretty solid as well.

I love those... Except when those stations start working badly and it's impossible to disable them (a month going where I live in London)

Where I live (Oslo, Norway), the predictions of Dark Sky and Weather Underground are both less accurate than just making an uneducated guess based on how the weather has been developing the last couple of days. I guess they work better in North America

In my part of North America, you can do better than any weather forecasts that I've ever seen by simply predicting that tomorrow's weather will be pretty much the same as today's. And all forecasts beyond three days are essentially worthless.

I was recommended yr.no many years ago and find it very accurate. I'm UK-based.

Fascinating - not one single overlap. Mine:

Firefox, K-9, Signal, KeepassAndroid, Syncthing, FBReader, Fast Notepad, OsmAnd+, Revolut, OpenVPN.

Mine: Firefox, WhatsApp (for family) & Telegram (friends and groups), andOTP instead of Google authenticator, New Pipe instead of YouTube, FBReader, Orgzly (notes), Phonograph for music and trying to move to LessPass for passwords.

andOTP is great, NewPipe is not as great (get YouTube Vanced, https://vanced.app/ instead). Why do you prefer FBReader over Moon+?

dang, went to find Dark Sky on App Store and it appears to not be available in Canada? sadness :(

Interesting... googled Dark Sky, tried out their web app, will probably download the mobile app. The original article's thesis seems to be on to something :)

BlackPlayer Ex (Music), MyBible, EasyWay (public transport), Multitran, WireGuard, Moon+ Reader, Smart AudioBook Player, CF.lumen, YouTube Vanced, Flud (torrent client), and YouTube Downloader (from XDA) are my essentials.

I like trying new apps, and will often go through a lot of apps to find ones I like.

I have Authy instead of Google Authenticator — not that there’s anything wrong with the latter, but if you lose your phone it’s much more of a hassle.

Apps I use at least several times a week (often multiple times a day). Leaving out some that I use but where there are likely many good options (e.g. TOTP authenticator apps, alarm clocks, calculators, etc.) and "any color as long as it's black" options like GMail and Youtube.

Apps of particular note that people could easily have missed are at the top of the list. Hope I kept lines short enough.

  - TripLog Mileage in plug-to-start mode;
  - Bouncer (auto-remove permissions from apps after you close them); 
  - DroidEdit Pro (multitab text and code editor); 
  - Join by Joaoapps (Pushbullet alternative);
  - Meteogram Pro from cloud3squared (Weather widget);
  - SMS Backup+ from Jan Berkel (pushes to GMail with label);
  - Nine (multi-account Exchange client); 

  - Firefox (plus "Dark Background and Light Text 0.6.10" and uBlock Origin);
  - Firefox Focus for untrusted links (no default browser set = always prompts);
  - Microsoft OneNote, Office Lens and OneDrive; 
  - PocketCasts;
  - Textra (for SMS) and Signal;
  - Nova Launcher Prime plus Will Windham's Vintage icon pack;
  - Swiftkey keyboard;
  - LastPass;
  - TimeClock Connect Pro from Spotlight Six;
  - Ultimate To-Do List from Custom Solutions, but I recently dropped ToodleDo so this may drop away.

And for frivolity, Pokemon Go and Calcy IV.

I too have a core set of apps (email, calendar, browser, phone, weather, sleep tracker, news, password manager, and note taking).

I'll buy a game once in a while (just got Civ 6 on the iPad and it's outstanding), but that's about it.

I figure most things don't need an app unless they are trying to take more data anyway. Anything telling me to install an app whose mobile experience is fine already is very fishy to me. I'm looking at you, Reddit.

Interesting, iOS has pretty locked down default settings and prompts you when an app tries to access most sensitive data. I think I’d generally prefer mobile apps from a privacy perspective because I don’t have to worry about things like cookies as much.

That's a good point, and maybe my hesitance to use these apps comes from ignorance about what data they have access to. It's just them being so pushy about it really makes me skeptical. Perhaps users are less likely to leave Reddit if they are viewing from the app? Richer ad content? More ad content? Notifications?

In a browser you have control about cookies, I'm sure iOS can leak more sensitive data even with default settings.

Android does that too now but many apps will just refuse to work until you accept.

I have to disagree with you here, even from a social media perspective. Namely, with apps that leverage hardware at a high rate such as the camera features of Snapchat and Instagram.

I agree that reddit might not need an app in the same way, but I'm sure there are ways they could improve their user experience by further leveraging mobile hardware in ways that don't relate to tracking.

Reddit lives on ad revenue. Mobile web reddit is not in fact incredibly optimal which is why a bunch of third party apps exist. These third party apps make it impossible to show ads save for promoted posts which people don't seem to like. Thus the official reddit app.

I prefer "Reddit is Fun" which seems to load faster than mobile web reddit and notify me on comment replies.

A primary reason why I don't switch to the app version for websites I visit (like Reddit, occasionally) is that I miss the ad and tracking blocker features I get in my browser.

> Reddit lives on ad revenue. Mobile web reddit is not in fact incredibly optimal which is why a bunch of third party apps exist.

That's a problem that reddit creates for itself as it purposedly degrades the site's experience on mobile devices with tons of dark patterns to try to push users to install the company's official mobile app.

Both Reddit and Twitter are intentionally designed to be annoying, so as to encourage you to install their apps. So frustrating.

Twitter had this some problem and went with the same solution of cutting out 3rd party apps. This is the age old piracy debate relived. People aren't using 3rd party apps because they want to skip ads, they are using them because they provide better functionality/user experience. Companies stupidly allow a 3rd party to take over all of their market share on mobile, and then make up complaints when they finally get around to making their own app.

> Anything telling me to install an app whose mobile experience is fine already is very fishy to me. I'm looking at you, Reddit.

So much this.

To add insult to injury, reddit intentionally degrades the site's experience in mobile devices with tons of dark patterns pushing users to their shady mobile app.

I use a Firefox mobile extension which helps greatly (no Reddit mobile ads)

I've noticed this too. On top of that, they nag you consistently to install the app. They have an option to disable the nagging. However, even with this option disabled, they still nag you via ads and every time you refresh the page it blinks a distracting "use app" animation at the top right. Reddit is in a downward spiral and I have no problem discontinuing the use of a web site I have been using for over a decade. I've done it before.

>Anything telling me to install an app whose mobile experience is fine already is very fishy to me.

Facebook is my #1 example of this.

The app bloated up like it's primary purpose was to take up space on your phone. So I removed it and used the mobile website.

At first you had to refresh the page to get new messages. No big deal, but a bit annoying.

Then they updated and didn't even need to refresh the page to get an update to the thread you were in.

Then a few years back they decided you can't get messages in the mobile website, you must use their app. Later I learned about mbasic.facebook.com and have to switch to that when friends message me.

Them not allowing you to use Messenger in a browser is just so disgusting. It’s like bullying. I’m happy people are jumping that ship.

> I don't have time to research whether or not they are going to abuse my privacy or have some horrendous TOS.

Do you think that web apps are better in that regard?

Yes, drastically. Wikipedia and Hacker News don't have access to the contents of my phone or desktop, just as they don't have access to the contents of my home or mailbox.

Yes. I don't give them permission to run code on my device arbitrarily and perpetually.

Think of it like this: If Hacker News required a fat client to function on your desktop, would you actually be here at all?

While I agree on your point, your comment made me smile as I’m reading HN using a mobile app

I'm curious, did you start using HN with the mobile app or migrate to it after you realized the usefulness of the website?

I migrated to the mobile app after discovering the site. I find the app greatly increases the readability of the site by adding just a few more colors, some icons for particularly popular threads, and a palatable dark mode.

You could always use https://github.com/openstyles/stylus to customize HN looks like

Oh wow, this is fantastic, thanks! I wonder if I can run it on FF mobile to make the site more friendly.

EDIT: I can, awesome. If anyone has a good theme to suggest, that would be fantastic.

So nothing that couldn't be solved by a better mobile friendly web site?

Which app are you using? There are quite a few on the Apple store.

> Think of it like this: If Hacker News required a fat client to function on your desktop, would you actually be here at all?

HN essentially is a service that provides only a couple of text views to list and read submittions and their discussions, and requires zero processing or interaction. That's hardly a challenging problem that requires a fat client.

If however we were discussing an application that required significant data processing, access to your personal data, or even access to photo ir video input... You'd hardly be able to implement something with HTML+CSS.

Case in point: twitter is very usable as a website but instagram is not.

They wouldn’t have to change much about Instagram to make it into Twitter.

If you take away those features then you cease to have instagram.

Instagram users only use instagram because it provides access to those functionalities.

What functionality are you referring to? The only one I can think of is photo editing. It’s not that difficult to make a photo editor in the browser. The only problem is performance. It would’ve been harder when Instagram was originally launched.

Instagram is perfectly usable as a website, at least on an iPhone. I guess some features might be missing, but the core stuff is there: browsing, adding photo, filters, stories.

How do you post a photo on instagram in a web browser? I thought that was reserved for the app.

Add a file input -> the users taps it -> ios asks if they want to use an existing photo or take a new one -> take a new photo. Result on android may differ.

Interesting, they don’t seem to allow this on the desktop browser. Changing the user-agent yo a mobile browser exposes it. What a strange decision.

I don’t use Instagram so can’t check the code, but I would guess they use something like

    <input type="file" accept="image/*" capture>

Which does what you’d expect on a mobile device (launch the camera) but behaves differently on desktops (opens a file upload dialog). They may have decided that this behavior would confuse users.

How do you ensure that a client only sees a picture/video once and can't save it?

You can't ensure this natively either.

The photos are available as publicly accessible jpgs through http. They’re not really trying to do that.

You can post a photo on instagram in a browser on iphone. you also can on other platforms by changing your useragent to webkit

> That's hardly a challenging problem that requires a fat client.

That's irrelevant. It doesn't matter that making a fat client for HN isn't necessary; that's completely beside the point. You're looking for reasons to ignore the stated premise of an analogy rather than accepting that the premise would be true.

It's like this. Say you were beginning to explain how network services work with an anecdote: "Say you need to go to the market to get a carton of milk." Suddenly your listener stops you and says, "But I don't like milk."

If your response to the above paragraph is, "But I don't know how network services work," then, congratulations, you can look forward to an exciting career in either comedy or politics, depending on whether or not you were serious.

Can you please not post flamewar-style comments to HN, regardless of how wrong or provocative you find another comment? We're trying for better than this here.

https://news.ycombinator.com/newsguidelines.html

On mobile, you don't use a fat client for HN? I've even paid money for an app (MiniHack) to have a nicer experience on mobile.

Truth be told, a full blown mobile app isn't needed to tweak the UI or provide different views.

Given that all websites are running code on your device, that seems like a distinction with a very small difference. The question is what has the better sandbox to protect yourself. The web is definitely better but somehow I still get tracked across sites and shown "relevant" ads. Mobile apps are also fairly well sandboxed and yet some apps that need your permission to do something useful also use that permission to do evil.

As web applications get more powerful, they will become a greater and greater source of the issues that currently plague mobile apps.

Most mobile apps are really just glorified websites that don't need anything above and beyond the web sandbox. But they're going to ask for those permissions anyway, because when an app is the only way to use a sufficiently popular service, people will grant them.

It is much easier when being nice is enforced by a third party - in other words, the webpage might have an interest in taking all your cycles, but the browser app has an interest in the opposite (battery life and whatnot). So far, this seems to work well - for all the gripes of FB Messenger taking all the CPU and requiring every permission in the known universe, the FB mobile web gets adequately sandboxed by the browser.

(I am aware that the Android app model has also promised some sandboxing, but apparently even in a low-permission mode, the protection seems to be rather anemic)

On the web I can install ublock origin, privacybadger, and a vpn. Even if you ignore the phone's personal data aspect that's a 99% improvement over mobile.

I can easily put an anti-tracking filter on my phone's web browser. Can't easily do the thing for apps.

> I don't even bother installing apps anymore, I don't have time to research whether or not they are going to abuse my privacy or have some horrendous TOS.

I'm going out on a limb but I'd wager you're not exactly representative of the general userbase for which apps are developed.

I would take that even further and venture a guess that 99% of HN readers are not representative of the intended general userbase of 99% of the apps for ios/android.

Also called the Malkovich Bias.

I disagree with the 99% estimate, I too take the same approach, and recommend the same to those who will listen. I would venture that number to be at least 80%

So you are saying that out of the 3B smart phone users in the world 600M will never install a app ever because they fear the abuse of their privacy or some horrendous TOS?

Yeah, no!

I think your intuitive judgement failed you here

The claim isn't never--it's for installing a small set of apps that you don't change for more or less the entire lifetime of your device.

Yes. Lurking around the app store of your device and installing random apps is a behavior for first-time users of the platform. Back during 2010 to 2015, almost everyone was a first-time user of the platform. From now on, downloading random apps is a job for ever younger audiences.

Crikey. I'm 48, a developer and I still install random apps.

Yes, I see and it's not binary like there's people who download random apps or not and there's people who are young or not, it's just that the numbers are clearly going in specific directions really fast.

I just bought a new iPad and one of the first things I looked for was apps that utilize the AR frameworks etc. if you are just gonna stick to the web it’s hard to take advantage the full power of your new device.

Yes because it's new and awesome and I'll probably do the same when I get an AR compatible device but it's not the reason we get the device. We get the device to install the apps we already like because we have stuff to do that we already started. Mostly. But especially for older audiences.

Older audiences tend to get replaced by younger ones.

Perhaps not privacy but the average user might be aware of losing a little battery life for each new installed app.

Luckily this is really not a problem in practice on iOS. Main reason I switched after years of frustration with Android. I use a ton of apps and the worst offender for background use is Hangouts of all things (I guess not surprisingly). So I guess I’m giving up a little bit of battery life but it’s really just one poorly designed app (I do not use hangouts nearly enough to justify its share of background draw)

A non trivial number of people don’t realize you can install apps. I know several non tech 65+ iPhone users the vast majority don’t use any app not installed on the phone.

Young kids seem to install apps all the time, but parents quickly learn not to let them spend any money on in app purchasing.

>A non trivial number of people don’t realize you can install apps. I know several non tech 65+ iPhone users the vast majority don’t use any app not installed on the phone.

Those don't matter much for the concerns of TFA, as they're unlikely to use some new fangled web app either...

do you know the origin of this? I suppose from the film "being john Malkovich"...

My personal guess is even more far-fetched but I think the ratio between useful and useless native apps reached 1/100 very, VERY fast after the industry successfully implemented wiring money as a feature. That was the one thing, even bigger than successfully implementing ads.

> 99% of HN readers are not representative of the intended general userbase of 99% of the apps for ios/android.

Not that I agreed with the website vs. app debate, but with all respect, that "userbase of 99% of the apps" was likely made up by those people who would also begs for bigger keyboard on their phones so they can send SMS a bit faster back in 2006.

This comment implies (like Fords: "If I had asked my customers what they wanted, they'd say better horses") that this 99% of customers are some backwards dwelling people who don't know the way forward (in your example, the no-keyboard touch screen).

But webapps are not some novel development they're not aware of. They are what existed before apps and during apps, and exists still, and people still spend most of their time on mobile apps.

People will spend their time on anything they feel useful to them, not the "mobile app".

I think you've been mislead by the current situation where everybody is using apps on their phone. But really, the "mobile app" it's just another thing that is attractive to them.

If you lock somebody in a room with only a TV inside, eventually that person will turn on the TV and start watching, even the TV only plays China Central Television channel one (FYI: It's boring like hell).

It's basically the same effect, the only twist here is that people chose to be addicted to their phones.

Don't let that effect blind you :)

As a side note, 99% of the statistics are made up, including this very sentence.

And 9 out of 10 concerns are unfounded, so don't worry about it.

On the other hand HN users influence is much greater than their numbers because many of them are doing tech support for family and tech advice for friends.

I used to do that a bit in the past, but not even my completely technologically illiterate father asks me anything anymore since he got an iPad.

While he is not the type to install new apps, all my friends are, and never ask me for any advice.

Anecdotal, but I think HN users do not have all the power you think they have.

You could also say their influence is greater because they are the ones writing the apps in the first place.

A tiny fraction of the people writing apps.

True, but that implies that someone here is after the money of these users, rather than educating them and providing them with the best (safest) options.

HN users represent < 1% of the userbase. But people often ask their opinions on tech products, so it is hard to claim whether they matter or not.

Do they ask their opinion? No one asks my opinion on software/hardware and I'm the only software engineer anyone in my milieu knows at all. But even setting aside how we evaluate the anecdotes that usually warrant that claim, as far as impact goes, HN users generally sport middle class backgrounds or better; the larger population has neither met nor really considers the opinions of a software engineer and typically mobile apps, and ones which don't have their user's best interests in mind, are lowest common denominator for a target since they're interested in reaping as much information as possible. It seems easier for a layman to just install whatever apps on their phone than to consider fielding an educated opinion. HN users may just like the idea of appearing important

If you slice the demographic data finely enough, everyone is a member of a tiny minority.

My non-technical retired mom also only uses about ten apps, for basically the same reasons, although she doesn't have the vocabulary and jargon we have to succinctly explain the same concepts. AFAIK its the same for my wife, aunt, and sister (different people, LOL) who mostly use Facebook app, being middle aged women.

I think it's representative of the average user a couple years down the road, as is often the case.

My grandma stopped installing apps for the same reason. Consumer behavior is changing.

well according to the article Google Store app numbers peaked in early 2018 so it might not be as unrepresentative as you think. Time is a zero sum game and there are only so many apps all of us can install no matter how enthusiastic we are.

I was curious about that so I followed through the link and saw there was a jump down of a million, and upon googling it, it seems that Google did a purge this spring. It doesn't have to do with users or developers but just Google doing a clean-up.

https://bgr.com/2017/02/09/google-play-store-app-privacy-pur...

Most people don't know how to unsubscribe from mailing lists... that have mandatory unsubscribe footers.

HN users are far from representative.

A few years ago, Flipkart was the leading e-commerce portal in India, and Amazon was a poor third, behind Snapdeal[1]. At some point in March 2015, Flipkart (and their fashion portal, Myntra) decided to go "app-only" on mobile[2], i.e., if you went to their website using a mobile browser, you would get shown a large interstitial to install their app, and nothing else. There was no option to proceed with their mobile website. Flipkart also gave away exclusive discounts for people downloading their app.

People rebelled. A lot of people rebelled. No one wanted to install their app to comparison shop. People just went to the Amazon site and bought whatever they wanted, knowing that at most, there would be a minor difference in price. Shopping on Amazon's site using a mobile browser was a breeze compared to Flipkart. The whole app-only strategy was a disaster.[3]

Flipkart admitted their mistake[4], and decided to re-instate their mobile website in November the same year. But the damage had been done. By August 2016, Flipkart had lost its leading place in the Indian e-commerce space to Amazon[5]. They never recovered, and to this day, they're still playing catch-up.

Not everything in the entire debacle can be attributed to Flipkart's app-only strategy. Undoubtedly, there are others aspects - like Amazon's prime being better than Flipkart's version, etc. But the app-only strategy definitely contributed significantly to Flipkart's uncrowning, and provided Amazon the much-needed entry point to becoming the market leader.

1. https://www.business-standard.com/article/companies/flipkart...

2. https://www.livemint.com/Industry/J9VeQxowSOlHU8ZMUParUL/Fli...

3. https://www.techinasia.com/flipkart-myntra-app-only-disaster

4. https://trak.in/tags/business/2015/11/10/flipkart-u-turn-mob...

5. https://www.livemint.com/Companies/c6vY9ta120G7cwewrCg6bI/Ha...

Very interesting story. I must remember to ask my Indian colleagues about this

As a heavy user of Myntra's app, I also have to say that their app is one of the biggest reasons why I keep coming back to them. The app is wonderfully designed, highly functional, and fast. The UX is uniform and the filtering options are top notch. Far better than anything the competition can offer

On a positive note, imo they have the best e-commerce PWA now. I recommend taking a look on mobile.

The question is not whether or not HN users' decision processes aare representative, but whether or not the end results are.

App adoption and use, overall, is low.

Consumers Spend 85% Of Time On Smartphones In Apps, But Only 5 Apps See Heavy Use https://techcrunch.com/2015/06/22/consumers-spend-85-of-time...

App Download and Usage Statistics (2018)

The total number of mobile app downloads in 2017 – 197 billion (a forecast)

That's an average of about 50/user, with an 80%+ abandonment rate, and a median all but certainly far lower.

How many Android apps are there now? Well, by June of 2017 it reached 3 million Android app mark! The current rate of its growth is more than 1,300 apps a day.

This is not a good thing.

..despite the sea of choice for mobile apps available for both iOS and Android, in real life people tend to use on a daily basis only a few. Here is how much exactly – 10 apps a day on average or 30 apps on monthly basis.

http://www.businessofapps.com/data/app-statistics/

77 percent of users never use an app again 72 hours after installing

https://www.androidauthority.com/77-percent-users-dont-use-a...

How Many Apps Do Smartphone Owners Use? Most apps are not even retained for a full day

A Localytics survey, conducted by Research Now in October 2015, reports that 49% of US smartphone app users use six to 10 smartphone apps each week.

https://www.emarketer.com/Article/How-Many-Apps-Do-Smartphon...

New data shows losing 80% of mobile users is normal, and why the best apps do better

https://andrewchen.co/new-data-shows-why-losing-80-of-your-m...

The more time I spend working with studies and research data, the more I'm starting to realize that a lot of it [numbers] is complete and utter nonsense.

197 billion app downloads? You don't have to be a researcher to know that the number is a "little" far-fetched. And maybe I'm delusional to the fact that might be possible. After all, I have only ever used Android and only with the default apps it comes packaged with. Other than the exception for WhatsApp and Messenger.

But honestly, in the markets that I work with, I see such blown up statistics that it makes me throw up on the inside. E.g. In 2018, 20% of all web searches are done using voice (assistant, Siri, smart speakers, etc.), and by 2020 that number is "going to be 50%".

Holy macaroni... I can already picture the dystopian reality where people walk around airports all talking to their phones just to look something up.

All that aside, mobile apps suck! I prefer a well-designed mobile website over an app at any time of the day.

>197 billion app downloads? You don't have to be a researcher to know that the number is a "little" far-fetched.

Is it? I'm a highly technical user, CS degree and all, 20 years experience, and I still downloaded around 20-25 apps last year. Of those, I kept like 4-5 on the phone, but the downloads are still there.

~3 billion * 20 = 60 billion app downloads already. And younger people are not as mission driven ("need to find an app for a specific task") and picky as me. Add to that casual apps and games, where people can download a new one every week (I rarely play games).

>But honestly, in the markets that I work with, I see such blown up statistics that it makes me throw up on the inside. E.g. In 2018, 20% of all web searches are done using voice (assistant, Siri, smart speakers, etc.), and by 2020 that number is "going to be 50%".

Yeah, that sounds like just BS PR from from SEO article pushing for some voice related product or service. Absolutely there are those too.

Well, I have over 40 non-Google apps, and I had to download them all again when I got a new phone last year (which I rarely do, but many switch every couple of years). And I probably downloaded and uninstalled 8 or 10 more - when I want do do something new, I often try a bunch of apps before settling on one. Also, some apps follow the model of using a "virtual app" as a key to unlock paid features, so that's two downloads for a single app.

Every update to an app also counts as a new download.

I think I never had more than 10 apps installed on a smartphone since like 2008. Not for privacy or security concerns, I just never had any use for more than a few core apps and always used the browser for the rest.

This doesn't express how a mobile app is essentially worse than a website - it only highlights how apps devs are doing it wrong and how mobile platforms vendors don't do their job preventing this.

Apps vendors should be discouraged from abusing access permissions, users should be warned if they probably do (in a visible and practical way) and be allowed to control them.

When a user clicks to install an app the system shows what rights does the app want and only gives choice to allow everything or just cancel. Instead it should disable everything by default, let the user turn particular permissions on explicitly and still install the app even if the user won't allow anything. It's the app vendor job to handle the cases when app can't access something.

As for TOSes - I doubt I understand why these should even be allowed by a platform. All the reasonable TOS terms are obvious and can be implied: the user can use the app for whatever is not illegal, the vendor can use whatever data the user enters the ways actually needed for the app to do its direct job (+ in non-personalized statistics calculation perhaps) and no other way.

> Instead it should disable everything by default, let the user turn particular permissions on explicitly

That would be great for power users but it wouldn't work for "the average dumb user" since they would just enable every permission when asked to do so if they don't know what it means and how it could be abused. And power users already have that in Android forks designed for them. (I can't speak for iOS.)

> All the reasonable TOS terms are obvious and can be implied

That's your definition of reasonable. With a rule like this a platform would kill off both freedom- and privacy-minded software as well as spyware (the general trackery), thus pissing off almost everyone who cares about those kinda things.

I am very, very sparing about installing apps. However, I'm even MORE cautious about browsing to websites, unless I can disable all client-side scripting.

At least I can firewall apps to mitigate any abuse they may engage in. That's much harder to do with websites.

Sounds like a miserable way to browse the web. Why not just isolate your browsing in a VM and trash it when you’re done?

I’m not security expert but seems to me if you clicked on something malicious, it costs you nothing as you’re not really exposing anything. Obviously don’t do your banking, etc on the same VM then go clicking around warez sites

It's not miserable at all -- in fact, I personally think that it makes most websites better. If a website can't operate without JS, then I just don't use that site. No loss to me.

VMs are another reasonable approach. I don't do that, though, because it's more hassle than it's worth to me.

> if you clicked on something malicious

That's not my primary concern. My primary concern is really to stop bad behavior that is very common across all websites (tracking, etc.)

just use Umatrix - https://github.com/gorhill/uMatrix

you can get the add-on for Firefox mobile also. It can be a hassle for the lazy user, but the addon will provide visibility into scripts, cookies and services used by websites.

Most scripts are blocked by default which dramatically improves page load times and privacy.

Yep. I keep <15 apps on my phone.

Given the fact that most people are already maxed out on apps on their device with just things like facebook and youtube, you're going to be hard pressed to get people to install your app just to try it out. With a progressive web app if they like your app they can add it to their device home screen without installing via the app store and without using up space on their device. Also don't need to pay Apple $100 a year but you have the same benefits as a native app. Now with wasm you can even include native performance from a progressive web app.

On iOS you won’t get notifications or be able to run anything in the background.

But if you are trying to start a business and the $100 a year you would have to pay to distribute your app will make or break you, you have bigger issues.

Too many applications abuse notifications and background processing anyway. If you need those features, you have to convince of far more value before I'll install a thing.

If you want an iOS app not to “abuse” background processing, there is a setting that will disable the ability on an app by app basis. Of course you have granular control over notifications and how they are presented on a per app basis.

I'd wager you are not a great representation of the market that 99% of apps serve. Apps abuse those features because it works well for their business.

>Apps abuse those features because it works well for their business.

In the short-run anyway...

This is true.

I allow certain apps to use notifications because there are things that I want to be notified about. My card balance from the Starbucks app, or breaking news from the New York Times app.

Both of those companies started abusing notifications. Their apps are no longer on my phones.

> I'd wager you are not a great representation of the market that 99% of apps serve

I'd wager no one is. Snapchat has 200 million users. There's some 4 billion people on the internet. Niches are niches, and no one serves everyone.

I wish "no one" was my customer then I could get a $40B IPO serving everyone :)

I think the point stands, and the article explicitely says "start with". Apps can come later in lots of cases. A personal example - I was hired to build the MVP Android app for a new greeting card company. 1 year later: It's been done for 6 months, no one's using it, no one wants to use it, everyone's using the web app.

On iOS you won’t get notifications or be able to run anything in the background.

Ok, but what are the downsides?

As a user notifications are a problem, but as an app developer they absolutely work to drive engagement, which is why so many apps use them. So if you’re trying to run a business that’s a pretty important thing.

> As a user notifications are a problem, but as an app developer...

So do you develop apps but not use them? I can't think of a way the hypocrisy isn't shocking here. "Do unto others" and all that.

I’ve developed software all of my career that I personally have no use for.

I develop software to pay my bills. We are talking about notifications, not killing puppies.

We're not talking about whether or not you are killing puppies -- I don't think you're a sadist. Just "make stuff you would want to use." Or "do unto others as you would have them do unto you." Or, if you prefer, "act only according to that maxim whereby you can, at the same time, will that it should become a universal law."

> Just "make stuff you would want to use."

So I should quit my job developing enterprise software because I don't personally have a use for enterprise software?

You think notifications are universally "a problem"? I find many notifications to be quite useful. Appointment reminders are VERY important to me and not at all a problem.

They probably mean generally a problem. Things like reminders seems like a good use case, but with all the apps trying to increase engagement, it seems necessary to mute notifications to an app right after downloading it.

You don’t need to “mute” them, just don’t approve the notifications permission request. Out of the 100+ apps I have installed, I have zero that annoy me with notifications.

Notification management is the problem. Not notifications themselves.

One part is on the system to provide the right management tools, the other is on the user to tell the system what notifications are important.

> Notification management is the problem. Not notifications themselves.

Notification abuse is the problem. Notifications have been so heavily abused by apps (and increasingly now by websites themselves) that makes the notification system itself look terrible.

Users don’t care about “driving engagement”. If your product’s user experience is important, or if is made with any empathy at all for the user, it won’t be spamming him with notifications purely to juice a startup vanity metric.

Unless I absolutely have to have the app (work or something), I avoid it. Webpage any day. It wont bug me in the middle of the night.

  3AM. Sleeping
  Phone: BING!!!
  Me: *groggy* shit, hope everything is ok...
  Me: *picks up phone*
  Phone: Rayman Adventures has updated!
  Me: *uninstalls Rayman Adventures, angrily goes back to sleep*

That’s an easy one - set Do Not Disturb to start at 10:00 and end at 7AM. It will still allow messages through for contacts that you specify.

In iOS 12, you can turn off notifications directly from the Notification Center for an app that bugs you.

Not to sound condescending but why would you have your notifications on at night? 10pm-7:30am do not disturb has been a godsend.

Work. Slack, texts, calls (could be anyone in my department whose number I don't know), PagerDuty, the random number PagerDuty call from. More of a pain to whitelist than to just avoid random apps with notification.

You have to specifically allow notifications after you install and launch the app. Just say “no” when it asks you for permission to send you push notifications.

Good luck, it sounds like a nightmare to be on call all the time.

One week out of five (rotation). Our software has to work (and work well) 24/7/365.

Service workers allow for push notifications and other background tasks.

iOS doesn’t support push notifications for web pages.

Are Safari push notifications desktop only?

https://developer.apple.com/notifications/safari-push-notifi...

Yep, desktop only

https://www.izooto.com/blog/ios-push-notifications-safari

And this might be why....

Web Push Notifications have clearly made difference in every marketing campaign it gets involved in.

Using email as a notification service works in this case. The user probably gets a notification for new emails.. at least if they do they chose to. :)

Email delivery has its own set of problems - mostly spam filters.

Even when the email does reach the end user, most people get so many emails it’s easy to get lost in the noise.

Using notifications to bypass spam filters speaks to the quality of the notifications you're sending.

If your e-mail is important, people will find it.

If your e-mail gets lost in the noise, then notification is probably not essential.

Getting caught in spam filters has little to do with whether the email is legitimate - just whether it is seen as a bulk email.

It does have issues when it comes to getting users to do something they don't want to do, but it is more than good enough to communicate with users about something they actively want to do.

I have no interest in the former.

With cordova you can create a thin wrapper that can add those features you need, starting with a PWA and extending.

To add a progressive web app to the home screen on iOS, doesn't the user have to use the share menu (and even then, it's buried a bit -- the user has to page the share options at least once to find it)? There's no way to display a prompt to add it, like with Chrome/Firefox on Android?

That's how it was the last time I checked, and it was enough for my employer to say no.

Yes. There is no way to prompt the user and, to make it worse, it can only be done from Safari. If the user is using a browser from the app store (like Chrome) it can't be done at all.

Yeah it is massively worse user experience than installing an app from the app store. I feel like anyone suggesting it is better has an agenda of some sort. It's objectively worse in every way for the user. Suggesting the $100 is a barrier to entry is absolutely laughable. If you are going to talk about money talk about the 30% cut they take. That's really the only reasonable argument in favor of them.

> Given the fact that most people are already maxed out on apps on their device with just things like facebook and youtube.

I don't even have those apps. If I did, my battery would be dead in a couple of hours. If I need to use those platforms, I do it through a web browser and usually I use a different browser (e.g. duckduckgo) isolated from my other browsing habits.

duckduckgo has a browser now?

Pick your poison: https://play.google.com/store/apps/details?id=com.duckduckgo...

https://itunes.apple.com/us/app/duckduckgo-privacy-browser/i...

I did not know they made one. Thanks!

Does Duckduckgo makes a browser now? Or do you mean that like most of non geeky folks you access websites not by typing an url in the address bar of a browser but instead by going through a search engine (in this instance Duckduckgo)?

>if they like your app they can add it to their device home screen without installing via the app store and without using up space on their device

As far as I'm concerned, until evidence to the contrary is provided, most people either don't do this or know how.

Relying on a pinned app for smart phones is a good way to kill your startup.

>you're going to be hard pressed to get people to install your app just to try it out

In fact, the statistics prove[1] that this is false. Most app installs are to try it out, and promptly delete it. Most users will delete your app shortly after installing it.

Informed opinions are nice to share, uninformed ones not so much.

[1]https://www.emarketer.com/content/most-apps-get-deleted-with...

> Most users will delete your app shortly after installing it.

Which is why it is so crucial to focus on the core utility of the app. If the app is fun to use, great. But the apps that stick around provide services that are essential to the user.

> >you're going to be hard pressed to get people to install your app just to try it out

> In fact, the statistics prove[1] that this is false. Most app installs are to try it out, and promptly delete it. Most users will delete your app shortly after installing it.

I'm really not following how this disproves what was being said. They were saying it's difficult to get people to even install for a test run, and you're saying most installs are test runs. There's nothing mutually exclusive about those two statements.

> Informed opinions are nice to share, uninformed ones not so much.

What a pompous response to make when you're not actually falsifying anything in their post.

>They were saying it's difficult to get people to even install for a test run

How is it difficult to get someone to install something for a test run when like 90% of app installs are test runs? So you're saying it's difficult to get someone to install your app at all, and that has nothing to do with your product or your marketing, but it has something to do with the technology?

And the PWA experience of pinning things to your home screen, which we have absolutely no reason to believe is widely done, is a better option?

There are 200 billion some odd app installs per year (primarily for test runs), there are ??? mobile PWA app home screen pins per year.

I'd like to see you try and support OPs claim rather than argue about my comment.

> How is it difficult to get someone to install something for a test run when like 90% of app installs are test runs?

The link you shared does not provide any statistics about the amount of installs relative to didn't-installs. It only deals with the set of people who have already installed, which says nothing to prove or disprove the point in contention (that increasing the set of people who have installed at all is difficult).

> So you're saying it's difficult to get someone to install your app at all

Yes [they are].

> and that has nothing to do with your product or your marketing, but it has something to do with the technology?

Based only on what they wrote, not necessarily. They only seem to be saying that the statistics you provided do not support your conclusion:

>> you're going to be hard pressed to get people to install your app just to try it out

> In fact, the statistics prove[1] that this is false. Most app installs are to try it out, and promptly delete it. Most users will delete your app shortly after installing it.

Ah, if I had seen you post this before I did, I wouldn't have bothered to post my own. Thank you for summing it up so succinctly and respectfully.

To recap. The claim is that PWAs are preferable to apps because "people don't test drive apps."

In fact, people test drive apps as a rule, and none of them install PWAs. I have shown a 2 second googled piece of evidence showing the degree to which people test drive apps. No evidence to support the ease of use or frequency of PWA homescreen pins exists/has been provided.

I'm going to stop responding to this thread unless your comment contains a substantive argument supporting OP.

> To recap. The claim is that PWAs are preferable to apps because "people don't test drive apps."

Ah. Now I see where the confusion lies. I don't think that's what the OP was saying.

> Given the fact that most people are already maxed out on apps on their device with just things like facebook and youtube, you're going to be hard pressed to get people to install your app just to try it out.

No one in this thread is claiming that people who install apps don't test drive them. Your evidence proves that they do, and no one has challenged or disagreed with that evidence.

The claim is not related to how often people who download an app test it out and delete it soon after. You're focused on the wrong detail. The claim is that they aren't even downloading apps to test them in the first place.

> I'm going to stop responding to this thread unless your comment contains a substantive argument supporting OP.

The OP is the only one arguing OP's perspective right now. Everyone else is just trying to get you to understand what the OP is actually saying rather than what you claim they're saying.

Let's try an analogy:

> Millennials are increasingly choosing not to eat at casual dining chains such as Applebees.

Responding to that by saying that Applebees is the most popular casual dining chain by a given metric does not disprove this claim. Whether people choose Applebees more often than Chilis has nothing to do with the fact that they're both losing millennial customers.

>They're saying it's difficult to get people to install an app even to test drive them, and saying it's easier to get them to visit your webpage.

But getting people to visit your website being easy does not mean that is where you should host your app. I don't believe people are going to say "Wow this website was useful, let me pin it to my screen." I believe that is a niche thing that nerds do, and has awful results for the rest.

>The claim is that they aren't even downloading apps to test them in the first place.

But there is nothing here to show that a PWA is the solution to get someone to download the app! The assertion is that the download is the issue, which there isn't any evidence for. If downloading apps was such a problem, why would people constantly be test driving apps? They may uninstall because of space or tracking concerns, or maybe it's a cleanliness thing, but I am refuting the base claim that "space on your phone" being a deterrent means that PWA is a solution.

Get the download through marketing, hold the download through value.

>Everyone else is just trying to get you to understand what the OP is actually saying

I think I understand what OP is saying just fine.

> How is it difficult to get someone to install something for a test run when like 90% of app installs are test runs?

As I said in my previous post I'm clearly missing something here. Why are those mutually exclusive at all?

You're saying that most installations are test runs. That implies people are leery about leaving an app they don't want installed. Why does that mean that it's somehow easy to get users to install in the first place? It sounds like they don't want a bunch of apps installed.

> So you're saying it's difficult to get someone to install your app at all, and that has nothing to do with your product or your marketing, but it has something to do with the technology?

No? When did I say your product or marketing were unrelated to install rate?

> And the PWA experience of pinning things to your home screen, which we have absolutely no reason to believe is widely done, is a better option?

When did I make that claim?

> I'd like to see you try and support OPs claim rather than argue about my comment.

Why does disagreeing with you mean I support the OP? Whether or not I support PWAs, native applications, both, neither, or even "native" wrappers around a PWA is irrelevant.

My stance is that your statistics don't disprove the claim made, and to then claim that the OP is arguing from a point of ignorance is not okay.

>As I said in my previous post I'm clearly missing something here. Why are those mutually exclusive at all?

If it was hard to get people to test drive your app, then 90% of app installs would not be test runs. They would be one time permanent installs.

>It sounds like they don't want a bunch of apps installed.

It also sounds like people are constantly trying new apps, no? Precisely what OP said people do not like to do.

>No? When did I say your product or marketing were unrelated to install rate?

Again, OPs claim was that somehow PWAs would be a preferable medium to app installs. This implies that the medium is the problem, which I am saying is not the case. "The reason people aren't downloading my app is because it's in the app store instead of being a PWA" is essentially the claim.

>Whether or not I support PWAs...is irrelevant. My stance is that your statistics don't disprove the claim made, and to then claim that the OP is arguing from a point of ignorance is not okay.

I think we both know that I could scour the internet and find 1,000 articles filled with statistics showing that people are constantly installing and uninstalling apps, and I would find almost none showing PWAs being successful in a business case. I'm not going to apologize for being incredulous about unsubstantiated conjecture that I have only ever seen proof of the opposite for.

>If it was hard to get people to test drive your app, then 90% of app installs would not be test runs. They would be one time permanent installs.

That doesn't logically follow.

Consider the possibility: Only 1 in 1000 of people who learn about your app is persuaded to install it. And then, of the small number who install it, 90% of those uninstall after 10 minutes.

I've exaggerated the numbers to make the point.

The point is it can be hard to get people to test drive your app and also have most installs be test runs.

They are not mutually exclusive.

> If it was hard to get people to test drive your app, then 90% of app installs would not be test runs. They would be one time permanent installs.

I think you and I have opposite understandings of the same data. You're looking at a high abandon rate after installation and claiming that it represents an increase in customers going out to find apps.

I don't think there's any substantive evidence of this. You have evidence showing that customers have a strict filter on what they keep on their phones. Why does a tightening of one stage of the pipeline have anything to do with a loosening at a previous stage?

There aren't a fixed number of installations that stick, uninstalling an application does not mean the customer will go right back out and install something else.

> I think we both know that I could scour the internet and find 1,000 articles filled with statistics showing that people are constantly installing and uninstalling apps

That people uninstall most apps they eventually do install does not imply a regular stream of new installations.

I wonder what would happen if you provided instructions for pinning a web app on popular platforms. Would users ignore this new information?

Many websites in the past have provided instructions for bookmarking.

We've been hearing the same pitch for literally years. Heck, I've made some of them myself: "This is it, people! Native apps have been rendered unnecessary!"

No PWA competes with a native experience. Not performance wise, not usability wise, and ultimately for the developer not even development wise. It may still make a lot of sense, and there are a lot of arguments for web apps, but the enthusiasm in this discussion seems disconnected from reality.

Indeed, right now we're seeing a big uptick in Instant Apps on Android -- go to a webpage and it actually loads a native app -- and I fully expect the same to appear on iOS.

There is a bit of app exhaustion, though I'd say it's much more significant on Android where users have been taught that it's user beware. It certainly isn't a technical limit, though.

> (paraphrasing) No PWA competes with a native experience even development wise

I'm personally too spoiled by the React debugging experience, and I'm saying this as a developer who mainly does C#. On the other hand, I know many developers who can't stand a line of JavaScript. The answer is here a huge "it depends, really".

For those who can't stand javascript, typescript was made for you

Typescript is just a huge bandage. It doesn't change that JavaScript has horrible tooling that changes every full moon. Not to mention it still keeps many of JavaScript's horrible warts.

When you can't avoid Javascript (which is extremely likely now), Typescript is a huge boon. Not everyone has the luxury of working solely on the backend.

Which "warts" does Typescript keep from Javascript?

Lack of a decent stdlib is a huge one. I would rather opt for Clojurescript or Elm.

Have you recently used Javascript? I ask because it's evolved pretty fast in the last 3 years. Even the past 2 years seem like a blur to me. Before 2015 it was unusable for me for large app development. Typescript evolves just as fast since it's a superset of javascript. 2.x was when it was usable for me. With 3.x, I have even fewer complaints. imo both are improving way faster than I can complain. That said, I still can't go back to Javascript after using Typescript. With rxjs and other popular 3rd party libs, it's hard to notice any limits. I'm not in love with Javascript, but it's just hard not to work with it since it's become the franca lingua of programming. Everyone knows it because everyone probably has to deal with it sooner or later. The ecosystem is also really strong. I have yet to had re-invent any wheels, which is something that's hard to avoid when using tech with smaller communities.

Just a few days ago actually, and probably will again later this week. It still lacks a decent stdlib.

...and for those who want even more improvement over JS than just TS, there's Dart in the Flutter world.

Why Dart instead of Typescript? How well does it integrate with npm modules? i.e. do I have to reinvent a lot of wheels compared to using Typescript?

Bigger difference than that. Flutter has you writing in high-level components that become native draw calls using Skia. There's an in-progress back-end that outputs to HTML, Canvas, and CSS paint instead of Skia. You aren't going to be using npm much at all.

https://medium.com/flutter-io/hummingbird-building-flutter-f...

If you are using Dart without flutter (eg, overReact), then you're going to use npm a bit more (though dart has better builtin libs which helps a bit on that front). There is a generator that uses typescript definition files to generate dart interopt files.

https://github.com/dart-lang/js_facade_gen

Or Scala.js. Huge stdlib (the entire jvm PLUS js) ecosystem) and no language warts.

React Native gives you the best of both

Or the worst. You still have to install an app. It isn't native. Might be some cases where it makes sense but an spa might be easier for everyone.

What is the actual problem during debugging react?

> ultimately for the developer not even development wise

Ehhh I'm not sold on that one, as someone who has done a lot of web dev and then transferred to doing some iOS and Android work. The languages you can work in (Swift, Kotlin) are fantastic, and the APIs and frameworks are great. But XCode and Android Studio are a hot mess and have been for years. Plus, compilation delays every time you want to do so much as run a test, waiting for app store approval before you can push out a bug fix....

IMO the developer story for native apps isn't that great. The web has a lot of wrinkles too, but coding with Typescript using hot-reloading and instant deployment of code to users has a lot going for it.

> No PWA competes with a native experience.

It does in one key area: it loads instantly. I know Android Instant Apps are out there too, but having to persuade your user to go to an app store and download your app before they can do anything is a huge lift. The web always, always wins on that front.

> But XCode and Android Studio are a hot mess and have been for years

I can't speak to XCode, but I agree that Android Studio is terrible. Fortunately, you don't have to use it to develop Android apps.

> But XCode and Android Studio are a hot mess and have been for years

Not sure where this is coming from but Xcode has been working pretty well for most of us.

I hear iOS devs say that, but code editing in XCode for me is very uncomfortable compared to Webstorm, Visual Studio Code or Visual Studio. There are no tabs for open files to speak of (XCode opens a new XCode for their "tabs"), the UI is very cluttered and things are quite hard to find if you don't know where they are - and it doesn't help that the UI keeps shifting around in updates, making Googling stuff harder because a lot of images of things became deprecated.

> but the enthusiasm in this discussion seems disconnected from reality.

This isn't the OP's point. The point is there is more friction to download new apps today so coming out with a new native experience is going to much harder to do in today's world. It doesn't mean that this is the long term strategy.

I'm speaking to this discussion, not the submitted article. Sure, start with a web app (though it really depends on the target -- if the web app yields a sub-standard interface, performance, or utility it's a dead end) with hopes of building an app when you have an established base.

However in this discussion there are a lot of people who are arguing from the perspective of a world that doesn't exist, based upon the same "this changes everything" argument we've heard every year.

> I'm speaking to this discussion, not the submitted article.

Precisely my point. That's not what the OP was saying. Per their comment: to get people to install your app just to try it out.

> based upon the same "this changes everything" argument we've heard every year.

Did we read the same comment? Where did the OP write "this changes everything"?

You're using OP to mean "the root-level HN comment" where @endorphone thought you meant TFA.

I am, but they're both making the same point, so it's kinda moot.

On the contrary.

In spite of my preference for native apps, I only got payed to deliver mobile Web and hybrid apps.

The large majority of CRUD apps don't need native features.

Also signed PWAs have access to native APIs on UWP and ChromeOS.

It is only a matter of time until the Chrome team exposes the same capability on Android.

Well, the React Native experience is very similar to React for web. So I'm not sure the developer experience is THAT far off. Also, for line of business, mostly CRUD apps a PWA is probably a better place to start. You get iOS, Android, Mac, Linux and Windows in one place with one codebase, with no install.

Frankly, I'm far less likely to install any app than I am to use a web app. In fact it pisses me off when I can't do updates from my desktop. No, it isn't a technical limit, it's more of an I don't want more crap running in the background sending notifications. I disabled both FB and Pandora notifications because they were annoying me and many users don't even know how. I'd rather not even go there to begin with.

If I use a web app that is very useful, there are times I'll see if there's a native app for mobile. If there isn't, I'm okay with that.

Right now, I'm working on an application that is web based and is for desktop use. I'm able to use react + material-ui and a few other modules and it's been going very smoothly. I can't say I've ever had a better experience with desktop app development, or mobile. ymmv.

> Indeed, right now we're seeing a big uptick in Instant Apps on Android -- go to a webpage and it actually loads a native app -- and I fully expect the same to appear on iOS.

I'd never heard of this but I can guarantee I will immediately quit and never install an app that gets forced upon me via this method. This is like the terrible "Try our app!" web popups, only 10x worse.

I realize like most people on HN I'm hardly the norm when it comes to mobile users, but UGH.

I saw it for the first time last night. I think it was Vimeo. It was actually much better than the "try our app" popup, because it was android asking if i wanted to use the instant app, and when i said no it just went away.

That's more reasonable. I wonder if these work via Firefox? Maybe that's why I've never seen them.

I normally use firefox too, maybe I accidentally opened the link in chrome last night. I don't have my phone around to test. I just remember thinking "ahh that's kind of neat" before clicking no.

As long as it remembers no locally and just lets me use the web app forever after it's a godsend.

They should't be available via Firefox, just chrome/google search

Instant apps aren't installed, per se. They don't have homescreen icons, can't send notifications, access protected data, run background services, etc. However if you have an app that has a rich interface for some need, Google slices off the pertinent parts and can serve it up transparently.

I've been hearing the constant whining about non-native apps for years. Pack in it, you've lost.

Everyone is writing Electron and PWAs at this point. Nobody wants to train C++ developers for a minor speed boost consumers don't seem to care about. They have a full build pipeline and dozens of trained webdev engineers ready to go, why invest in a whole new product?

A LOT of people have looked at the cost/benefit of electron VS something like QT. Many have decided that the performance hit isn't worth the cost of developing the C++ build pipeline and training developers on how to build QT apps:

- Spotify

- Visual Studio Code

- Atom

- Invision

- Slack

- Mattermost

- etc.

The web tooling is actually pretty nice if you stay in your lanes. CSS / HTML / Javascript can create a really nice experience if you don't do a bunch of dumb stuff and make pragmatic decisions.

Wunderlist had/has native apps for all clients. One of the main reasons people LOVED the app. It's certainly not a guarantee, but if you want that final level of polish for a consumer app, native is the way to go.

There has been a ton of whining over the years about native vs some JS framework and I'm tired of the shit too. Now I can't speak for your QT but I can for IOS. Do you know how much easier it is to learn swift + xcode vs learning html / css / js in a framework that compiles to native code? The answer is very much.

I think it depends a lot on what you mean by "compete". I'm a heavy Twitter user. I now only use its PWA on Android and it's fine. I wouldn't call it perfect, but it's certainly very usable, and I suspect most non-experts would have a hard time telling that it's not a native app.

Especially for MVPs, I think they're competitive for a wide range of uses. The UX may not be as polished, but that can be more than made up for by instant availability, instant updates, lower dev costs, and faster release cycles. The ability of a new company to learn is limited by release cycle time, and I think fast, low-cost learning is a huge advantage to startups.

>No PWA competes with a native experience. Not performance wise, not usability wise, and ultimately for the developer not even development wise.

How much of a usability advantage does an app for, say, IMDB offer that the regular website doesn't? Or the loyalty card apps that a lot of grocery stores and coffee shops use now?

A lot of apps on the market are functionally just wrappers put around a poorly optimized website. They'd be better off putting their efforts towards making a great experience on the mobile web instead of trend-chasing.

In the iOS environment Apple could actually help here through a simple awareness campaign. You can save website bookmarks as icons on your home screen as it is, but this functionality is little known and not easily discoverable.

Companies saddle their mobile web experience with excessive animations, gobs of dynamic elements, screen-hogging navigation bars/sharing buttons/overlays, poorly-implemented infinite scrolling, etc, etc... and then they conclude that mobile web sites are bad and native apps are good.

People have forgotten that you can build a website that isn't a user-hostile SPA train-wreck.

Amen. Twitter[1] and Starbucks[2] have both built gorgeous PWAs in the last couple of years that are perfect replacements for 98% of a native app's functionality. It can be done, people!

1. https://blog.twitter.com/engineering/en_us/topics/open-sourc...

2. https://twitter.com/davidbrunelle/status/905931990444244995

Done in-house, I'd bet, and that's another disconnect to overcome.

Outsourcing website and/or mobile dev to external company, the focus is often more going to be on whizbangflashy, vs sleek/slim/fast/simple, to justify whatever budget they're getting.

I do find that I end up using the mobile web client for twitter as often as i do the native client. It's not much in either case, but the mobile web client is generally 'good enough' such that if that's where I land, I don't feel a need to jump to the native client.

Twitter's mobile web site is the most unreliable site I encounter with any frequency. Fully 80% of the links I get to Twitter fail to load, telling me to try again, which only works when I fully refresh the page.

I don't know how it compares to their native app.

You're missing the point. A mobile app isn't for the benefit of the user, it's for the benefit of the analytics and marketing teams. They want to be able to know who's engaging with it, and spam them with notifications.

> No PWA competes with a native experience.

Sure they do. Consider Reddit's website (well, before the atrocious UX updates the past few months) vs. their mobile app. It work(s)/(ed) perfectly fine.

There's literally no reason to use their native app (well, other than their ridiculous, atrocious, never-ending prompts on their mobile website to use their native app).

Here's a tip - take any reddit URL and add '.compact' to the end. Example: https://www.reddit.com/r/all/ becomes https://www.reddit.com/r/all/.compact

This uses the original reddit mobile site. It is very fast and clean.

At least they are consistent, go to the compact site, giant banner appears:

> You've been invited to try out reddit's new mobile website! try reddit's mobile website

Luckily, it's easy to block with AdBlock Plus.

> well, other than their ridiculous, atrocious, never-ending prompts on their mobile website to use their native app

In your settings there's an option to tun this off (I agree, a truly terrible user-hostile default).

That's because Apple and Google are doing everything in their power to keep apps tied to their app stores. If they wanted to they could make progressive apps runs as smooth as native apps tomorrow by enabling better native integration, but they won't.

Google makes a ton more money off of web search than their app store and they keep releasing new features that allow you to access your devices sensors so I doubt that for google.

> No PWA competes with a native experience. Not performance wise, not usability wise, and ultimately for the developer not even development wise.

This would be correct, in theory, if companies only made apps where the extra power that the app has is necessary, and would stick to websites otherwise.

In practice, this isn't true, and anyone who has ever opened any major app store knows that many popular apps in it are, in reality, just wrappers around a webview, and the only reason why they are apps at all is because the management demanded that there is an app, like all the cool kids have these days. Most apps really would be better off as just websites precisely because they don't need anything that an app gives, but often suffer from what it takes (e.g. ad-hoc navigation versus standard browser controls).

> With a progressive web app if they like your app they can add it to their device home screen

Does anyone even know they can do that? I've never seen anyone do it "in the wild". Would love to see any stats around usage of pwa's added to home screen.

I run a PWA and can track this based on the "standalone" property. But what's more telling is customer requests. It's a confusing request that it seems like many potential customers haven't seen before. Lots of people ask me to "just put it in the app store" despite having an FAQ with instructions on how to add to the home screen. Many more say they will become customers "when you have an app"

Because of this, I'm working on making my UI more app like and using Cordova to publish it to app stores.

edit: My app is still pretty small but the percentage of logged in users who have ever loaded from the PWA is < 5%

Users always have to be educated, whether it's PWA or how to pronounce the @ symbol a mere 20 years ago.

Chrome on Android will show a banner to add it to your home screen.

True, the situation on Android is much better than iOS. Also after you do this, you get an icon in your app drawer and not just the homescreen.

iOS doesn’t have an app drawer, so I’m not quite sure where you’re going with your comparison…

Yep, sorry that was confusing. I was comparing to previous versions of Android. It's relatively new that it's added to the app drawer.

It also shows up as an app in the system menu, with natively controlled permissions, notifications, battery restrictions, etc.

I run a website where I get asked quite often to make it into a mobile app, and I reply by showing how to add it to the device's homescreen and not need an app. Never ever ever in the few dozen times I've had this conversation has anyone ever actually done it. In fact most responses to that suggestion have been overwhelmingly negative. They want an app, even though it'd just be a wrapper around the site with no added features, they want an app.

Bingo. That's exactly what end-consumers want. If you have a good enough PWA, you can wrap it in a single webview and push it to the appstore/play market. The presence is the important part, the actual implementation does not matter.

Why not just add the ability to list a PWA in the app store? At that point the installation is intentional and the overhead on Apple/Google is minimal... and Apple still gets their $99/year to allow making it easy to put your app's icon on a user's home screen.

On one hand I can see rational reasons for them not allowing it; Apple doesn't control when you push updates to your PWA, so they don't want to give their seal of approval by listing it in the app store.

But on the other hand, it's pretty depressing to give up the dream of instant app updates, by wrapping your shiny PWA in Cordova and promising to never execute external JavaScript.

I recently picked up app development and I went for React Native + Expo. I was shocked that I could push OTA updates without going through the store. Basically if my app was approved once, I can make huge sweeping fundamental changes to the app without running it past Apple again. Basically the only things you can't change are the permissions. Anything else is fair game. And with Expo it was basically trivial to get an app up and working (don't take this as biased praise in favor of RN or Expo, I still have serious dissatisfaction with both technologies, but this particular part was shockingly seamless).

So I guess moral of the story, wrap your PWA in an Expo app and update as often as you want without having to resubmit.

Oh, cool. I think I was confused about Apple's rules on this. I guess you are allowed to push minor updates to your app, as long as it's JavaScript being executed by their JavaScript engine/WebKit.

But apparently there is (or at one point was) a restriction that says you're not supposed to add new features via remote executable code pushes [0] (although I'm not sure if they'd notice).

[0] https://rollout.io/blog/updating-apps-without-app-store/

Pushing updates without going through the app store has been allowed by Apple as long as the assets being updated are not binary code; specifically CSS, HTML, and JavaScript are okay (I wonder whether WebAssembly will be considered "binary" or not... will be interesting to see). I convinced a former employer to abandon their Xamarin-built app and switch to Ionic (Cordova with nice, fancy UI additions) specifically because they could push updates to the app without going through the app store. Officially (unless things changed?) the Google Play store says this isn't allowed but there is no review process with Google Play so publishing a new app and then pushing an update for the previous version won't likely get you in trouble. If someone HAS been busted by the Google Play cops for this it would be news to me.

Ionic CEO here - have seen no issues with anyone using our deploy (web-content remote update) feature on either platform to date

Microsoft does this on UWP.

well i had a different experience. we dong have a native app, but some of our b2b customers wanted an "app" to use on mobile. The app is responsive so we show them how to add it to homescreen and that's it.

It might be business customers are more willing to jump through hoops. My website is targeted at the general public in a small midwestern town, where technology is a four letter word.

Indeed. Business customers aren't representative of the greater public because they tend to take their issues directly to IT support, they get it sorted, and don't think about it again. Your every day average user doesn't have an IT support department to which to go. They do, however, have the App Store/Play Store.

My guess is that it's rare. Hell, I didn't even know I could do it until I was developing a PWA. I've never seen a website that offers it.

It would be very nice if phones put that feature front-and-center. My guess is that they sideline it so the web can't compete with their native apps.

It really reeks of the same kind of anticompetitive behavior that got MS in trouble twenty years ago.

Apple is way worse in this regard, but Google is still incentivized to keep at least PWAs from being first class citizens on Android because they don't get their app store sales cuts or microtransaction money from them. And at least on Android you can install other browsers.

People consider the hegemony of Chrome to be the premier threat to the open web but I keep feeling iOS Safari is the real drag. Apple is so heavily incentivized to sabotage web tech that could encroach on their app store revenue.

They do it on ChromeOS though, so I guess it is politics as usual.

You can use browsers other than Safari on iOS.

All browsers on iOS are essentially wrappers for Safari. They all have to use the built-in iOS WebView; per Apple's rules you're not allowed to execute JS or anything like that on your own.

Which means no extensions for iOS Firefox nor Chrome.

I can’t install any adblockers on my iPhone, the best is Safari which has some built in anti-tracking but that’s it.

I wonder if there is anything like a PWA store?

Win10 store allows PWAs, and even automatically lists the ones it finds online (via Bing crawler).

https://developer.microsoft.com/en-us/windows/pwa

I find it all oddly ironic. Remember back in late 90s - early 00s, Microsoft was generally hostile towards open web and the (then only just appearing) web apps, because that undermined its desktop app monopoly? Here we are 20 years later, and now it's Apple and Google protecting their turf on mobile, while Microsoft is promoting PWAs to entice the mobile devs into supporting Win10.

I have a HN bookmark/icon on my home screen.

Yes. Most of my users are using the page as an app and the usage has only increased the last year's.

PWAs are becoming reasonable for mobile devices and could really help with some apps for initial distribution and usage.

I might implement the first version of mobile reading within Polar (https://getpolarized.io/) by using a PWA.

The general idea is to have both a PWA and a native app.

The PWA gets the users addicted and then you can have an 'install app' button within the PWA.

If your app is insanely complicated it's possible to have your app as just a PWA on steroids (though not always straight forward).

We're using Firebase (just wrote up a post about it here):

https://getpolarized.io/2019/01/03/building-cloud-sync-on-go...

... and the cool thing about Firebase is that there are SDKs for basically all platforms with really solid mobile support.

https://www.apple.com/newsroom/2019/01/app-store-caps-record...

The list of top grossing apps paints a rather different and rather unflattering picture of the app store economy.

https://www.apple.com/itunes/charts/top-grossing-apps/

I agree completely. What’s surprising is that Spotify is still a top grossing app even though they haven’t allowed new in app subscriptions for a year or two.

I guess those are legacy subscriptions.

But on the other hand, that also goes against the narrative that PWA’s will replace native apps anytime soon. Most of the apps on the list:

1. Require some type of DRM (streaming media)

2. Take advantage of in app consumables like games where the immediacy of being able to capture the whales through in app purchases make sense.

3. Need the performance of native apps.

After years of experience:

0) Native means you get notifications: APNS, GCM.

1) If you don't go native, you go SEO. SEO is hard.

Web has had notifications for years too.

Not on iOS. https://caniuse.com/#feat=notifications

I have never clicked yes to allow web notifications in my life. I have only once added a site as an icon on my mobile. The whole premise of this article is false. The real answer to whether to start with a site or an app is it depends. At Bibimapp we have an app [0] requiring sign up etc (launching beta next week) and a related site which is free to browse which provides value but approaching the problem space from a different angle. The site is essentially the top of the funnel.

[0] https://bibimapp.com/

"Web has had notifications for years too."

Yeah. But people don't use desktops anymore.

So what are all those laptops on every coffee shop I go to?

Native means you also get background processing and access to things like geolocation and camera feed w/o needing a plugin.

Can you provide data to substantiate your claims? I'd truly like to know if this is true, because my intuition and experience tell me the opposite. People like downloading and trying out new apps. I also disagree with the premise of this post. Many people exclusively use their phones and would never utilize a web version of an app.

I think it depends... it's probably bias to more technical users vs moms, dads and kids generally. I know I'm far less inclined to use a native app for something new. I also hate when I can't do updates from my desktop. I use linux, windows and mac regularly.

Of course via cordova, it's easy enough to create a thin wrapper and add a few bits for notifications and better capture integration (if you need photos/video, etc). Since I'm mostly doing react these days, I'd consider react-native a natural next step.

> Given the fact that most people are already maxed out on apps on their device

This argument has been given a few times here. But I question whether web applications are a solution. Aren't these just more applications that people are overwhelmed with but harder to find and use?

If you've got a brilliant new app (web or mobile) I'm still fatigued either way. The mental effort is the same. Saying the mobile apps are "maxed out" seems to me to apply to web apps in just the same way.

I think they were referring to RAM or storage space rather than human fatigue or mental effort, although I'm not a native English speaker so I may be misinterpreting.

Indeed, there are many low-end smartphones that run out of storage space after one or two years even without installing anything, just with the automatic updates of the preinstalled apps (Facebook is a major offender here, why does their app grow more and more?). Being forced to install an app to use a service when you are in this situation is highly frustrating.

What year are we in? I've got a $200 phone with 128gb storage and 4gb ram. It will be a long time before I can max this out if at all.

Not everyone buys a new phone frequently, and not everyone can or will spend $200 on a smartphone without a second thought, especially at the global scale.

Smartphones are getting better and hopefully this problem will be irrelevant in a few years, but I still know plenty of people for whom it's still very real at the moment.

Sure, I was part of a startup designing apps for emerging markets and I understand the restrictions but I think these restrictions changed somewhat for phones bought from 2018 on. I just returned from 2 weeks in India and it was common to see people in the streets with ~$200 phones

Sub $100 phones rule the streets here in Nigeria. In those cases storage is very precious and limited.

The flip side is that people who can only afford sub $100 phones are not as valuable targets to develop apps for.

I think PWAs are still considered more of a novelty.

If they really ever catch on, I could see where apple might require you to start hosting some kind of cross-signed cert or something that you can only get from Apple to install/use them.

> If they really ever catch on, I could see where apple might require you to start hosting some kind of cross-signed cert or something that you can only get from Apple to install/use them.

And you would find this acceptable? I would find this appalling. Apple can charge for the App Store because over the overhead. It hosts the servers that have to transfer the apps. It runs an App Store to help with discoverability. It also maintains the API developers use.

In the case of PWAs, none of these are true. There is no overhead for running an App Store for Apple. I could see there being a fee to be in the App Store. That should be up to the developer though. If you are doing the work of getting people to install the app yourself, you owe them nothing.

In Apple's defense, they do have to do the work of building in PWA support to iOS/Safari. But if you are paying over $1,000 for a device you better damn well be able to install a PWA.

Microsoft does it, if the PWAs are signed, they get full access to UWP APIs, just like any other native app.

That's only the case for useless apps. Try creating a useful one. For example, here's my old hobby project, now open source, that got more than 100k installs with zero marketing: https://github.com/Const-me/SkyFM

I've never added a PWA to my device. How do you find them? I suppose they don't get added to the store so no way to search for them there. Do they show some popup when you visit their site?

Is there some directory of PWAs?

What does "maxed out on apps on their device" mean?

> With a progressive web app if they like your app they can add it to their device home screen without installing via the app store and without using up space on their device.

Just no. As near as I can tell, PWAs are worse than apps from a security point of view. The odds approach zero that I'd ever dare to use one.

IMHO, the reasons mentioned are not very convincing. Your goal, when starting a product, is to find whether you serve a real user need or not. So, start with a website or a mobile app depending on what suits your users needs better. Where would a user naturally go (web or mobile) for that need? That's the platform you start with.

It takes longer to get to market with an app, especially if its your first. No contact with users == death.

But you also only get really once chance that with that user. Give them a crappy experience and they'll never come back. A web app is already starting at -100 points.

> Give them a crappy experience and they'll never come back.

Is that even true for a website though?

In the early days, that is actually okay. Unless you did a large marketing campaign (you shouldn't), the number of people who never come back because of your crappy MVP will be small. And the group will likely be small enough where your product doesn't get a bad reputation either. But you absolutely need to get feedback from those users who never come back so that eventually the product gets better and gets traction. I think tools to collect that indirect feedback (e.g. click heatmaps) are more mature on web overall. Collecting direct feedback, talking to users, is equally hard regardless of web or mobile.

...in absolutely no way is this true. Where did you learn this?

Who taught you that websites are 100 points worse (on whatever scale, you never say) than mobile apps?

Very frequently, the right answer for an MVP is something like "an e-mail newsletter" or "a Salesforce extension" or "a CSV file" or "an IDE plug-in" or "throwing a pizza party for your target market and performing the service for them yourself". I've seen startups charge $3K+/month to dump a CSV file on a client's FTP server. Go where the user can most conveniently make use of your product; for a lot of businesses, that is neither app nor website.

I agree. In many cases, don't start with either (app or site). In other cases (or in cases where you've grown up to the point of needing an app or site), start with one that is right for your user.

This is so right on. Yes there are pros/cons with PWAs vs native (and with the other simpler examples listed below. Put all those into the mix with questions about your riskiest product hypothesis, access to most representative users and pick the winner. Biasing always toward PWAs sounds more like a bias toward a particular solution yourself vs what makse sense for your business.

Having said all that, I agree that for many simple apps the better user experience, in-app, isn't that much different except in one major way: discover and re-discovery later. When you install an app on your phone, there is a nice big button on your screen at all times so you can come back to it. Getting people to install is hard but once they have it is way way better discovery experience than a web app. Adding a link to your desktop still feels pretty janky (right or wrong) to most people and most people don't know how to do that.

>An App Store developer license costs $99/year, whereas a Namecheap domain is less than $10/month.

"A developer license is $99/year, but a website is less than $120/year!"

Very weird statement.

Looks like a mistake, domains are $10/year

Maybe OP meant hosting.

If you use Lambda or some other auto-scaling thing, hosting is effectively free until you get traffic.

Hi there, apologies for replying to a completely irrelevant thread (HN locks down replies), but was curious if you are planning to Show HN your project [1]

[1] https://news.ycombinator.com/item?id=18611208

Hi - yes I am soon. Just got side tracked with holidays etc. I'm also in the middle of trying to find a buyer for my other side business. Flippa produces a lot of interest, not sure how many are just tire kickers though.

Cool, thanks for the reply.

Can I ask what you're selling on Flippa?

Even at $10 a year, it's a very weak statement to make about costs. If you can't afford to pay $90 more a year on something simple for your startup, you have other problems.

My thought as well. 90$ won't make or break someone's startup.

If the purpose was to make a point about app stores being walled gardens, then it could have been made more effectively. Edit: and you're going to be buying at least one domain name anyway, if you have a Serious Business.

The magnitude of that cost savings is insignificant compared to the cost savings realized in dev time and operational complexity. (Because you're probably building a HTTP backend, right?)

Also, even if the domain registration is $10/yr, it totally neglects hosting costs. The cost of dev license versus domain registration is a ridiculous factor when deciding a product strategy for your startup (either way).

The real answer to this question? (Surprise!) It depends.

I have worked on a side project for the last year or so that will have a website and a mobile application. It has social elements, so ultimately I decided to prioritize the mobile application, because that is where users have access to their contacts, and do most of their communication with their friends and family. There are, of course, other business ideas that would be better suited to a website (or electron app?): work-related tools, dashboards, etc.

A lot of people on here are talking about PWAs which are amazing, but iOS Safari is missing the Web Push API. Unfortunately, that makes the effort nearly futile because you have to build a separate mobile app anyway.

If you are an Apple engineer and can help get that shipped, I will buy you a nice dinner. If you're an Apple recruiter, hire me so I can help ship it! :)

> If you are an Apple engineer and can help get that shipped, I will buy you a nice dinner. If you're an Apple recruiter, hire me so I can help ship it! :)

Not that I know anything not known publicly but I suspect the reason Apple iOS Safari does not have Web Push API is not because they lack developer resources to write the code to get it done. Apple is under tremendous pressure to increase revenue from its "services" right now so it seems a bad time to strengthen PWAs.

Wouldn't you say though if that push notifications were allowed PWAs would cannibalize the revenue of the ios app store? It doesn't seem like apple would have much incentive to do this as PWAs would eventually compete with their own platform.

Apple makes most of its app revenue from in app purchases and most of that is for consumables for games.

The types of apps that could be a PWA wouldnt make Apple any money.

You can integrate Apple Pay pretty seamlessly into a PWA. It actually adds a lot in the way of feeling like a real native app.

I don't think that's really true, because Apple has strengthened PWAs in a lot of ways in recent iOS releases.

I personally suspect it has a lot to do with waking up a service worker outside of the browser lifecycle - given how aggressively iOS monitors apps no longer running in foreground I can imagine that being quite a hurdle to overcome (without compromising the reason for doing it in the first place).

> I personally suspect it has a lot to do with waking up a service worker outside of the browser lifecycle - given how aggressively iOS monitors apps no longer running in foreground I can imagine that being quite a hurdle to overcome (without compromising the reason for doing it in the first place).

I don't think this is true. Please correct me if I am wrong but there are ways to handle the scenario. Apple could create nsfirebase (basically an Apple version of Firebase for lack of a better name) and require all websites to talk to nsfirebase and the iOS device only needs to reach out to that particular end point on nsfirebase which minimizes battery loss. In addition, it could say something like your worker can only run in the background if a user has been to your site in the last n hours (where n is something Apple can monitor and update with any iOS release).

> In addition, it could say something like your worker can only run in the background if a user has been to your site in the last n hours

But iOS apps can't do that. The only way they can guarantee background execution is if they are playing audio, or say they're a VOIP app. Of course, Apple could change those rules for Safari, but what I'm saying is that I don't think it's as simple as just flipping a switch - it'll need OS level changes, which will slow things down a lot compared to stuff like offline caching, which the Safari team can implement themselves.

There's also a fundamental difference between iOS notifications and Web Push ones - the notification is already created in the iOS payload. So it doesn't need to wake the app up at all, it just shows the notification. That's not possible with Web Push, so it's a change compared to how iOS handles every other notification.

Getting background location updates will also let your app stay unsuspended. You better have a good reason for getting them, though, or your app won't make it into the store. (Ditto for VoIP or audio apps; the app store reviewers are wise to tricks for mimicking services in a non-service architecture.)

This is interesting:

> the app store reviewers are wise to tricks for mimicking services

Has that been written up anywhere? I'd really like to know more about the reasons behind this arms race.

>It depends.

Sadly I think to often people are to focused on "apps" fail to make that consideration. Businesses seems focus on getting their little icon on the home screen of people phone, to keep them coming back to their service. News sites are especially to blame here.

There are also some app that completely fail to recognise that there are people who need the service an app provides, but doesn't have a smartphone. For instance the only way I can buy a bus ticket with my debit card is to use an app on my phone, but what if I didn't have a smartphone?

Quite frankly not having a smart phone in 2019 is becoming not having a computer in 2000.

Yes you can live just fine without it but you're the exception not the rule.

"No one wants to download apps anymore" I stopped reading there. The major advantage of the App Store (primarily for consumer apps though) is the fact that there's some kind of filtering at place. If I find a quality app, I want to actually pay for it. In fact, if it's a good app - then I prefer to pay, as it often ensures some kind of continuity in the app's lifecycle.

What kind of filtering is that? The fact that there is (hopefully) no malware? - A browser should prevent that as well and has similar failure rate as app review. The app store doesn't tell me anything about whether it's useful or of high quality or anything.

From my perspective many apps are single tab browsers with additional spying capabilities.

For one, you have a chance of actually finding quality native apps. Web apps are not completely optimal no matter how you try to sell it. I have nothing specifically against hybrid or web apps (they have their place), but speaking from personal experience I've had most luck with apps that were (1) native (2) cost a few pennies. I say this wearing the consumer or end-user hat.

As a developer, I also prefer writing native apps (iOS and Swift), but I certainly understand the occasional disadvantages here too. This is in fact why I'm also developing a Flutter app at the moment - we needed a cross platform MVP but there's too little time and resources to develop for both platforms. But given enough time and resources (e.g. post MVP) I'd mostly go native.

>For one, you have a chance of actually finding quality native apps.

Really? I can't recall ever finding a useful app through the app store. It's so littered with junk apps and deceptive copycats that discovery is a miserable experience.

Seriously?!

Check out Things 3, Overcast, or Timepage (and Actions) by Moleskine. You don’t recall EVER finding a useful app?

If we follow your suggestions we're finding apps through third-party recommendations, not through an app store.

I have found useful apps through a store, but it does often involve sifting through a lot of misleading junk.

I can't recall ever finding a useful app through the app store. There are plenty of useful apps on my phone, but I discovered them through Google or personal recommendations.

The comment that sparked this whole thread was "The major advantage of the App Store (primarily for consumer apps though) is the fact that there's some kind of filtering at place". I don't think that's true - I think that the app stores are so littered with junk and offer so little curation that finding something on an app store is a fool's errand.

From a developer's perspective, I don't think that a native app offers any advantages in terms of discoverability or onboarding. The app stores are full of polished, well-designed apps with incredibly low download counts. A native app might be stickier than a web app, it might offer a better user experience in some cases, but installation presents a natural obstacle to previewing and sharing. I think that developing for the web should be the default choice unless you have a compelling justification for going native.

So you've found a quality apps by only finding it via the app store? It just sounds odd to me because I literally never have. The only way I've found them is by researching it on the web beforehand.

The top apps are still more or less the same staples as 5 years ago. While it’s harder to find a new excellent app, it still serves as a good introductory guide of what most people are going out of their way to install on their phones.

I've find apps through the App Store app and the web, or magazines etc. Even word of mouth. That's pretty common. You're taken through the App Store at some point to download the app though.

It is true that a native app can have a value due to integration into the system. However many apps aren't really good and built using cross platform toolkits as doing development two times for the two major platforms is expensive ... There are good and worthwhile apps, but it's low

the way you use your phone is different than others though. in the non-tech world people use less than 10 apps, email, social media, messenger, and maybe an investment portfolio

I remember trying to get people to download an app I made in college and it was pretty hard, even when the app would show up #1 in the app store results after searching the name. Website would have been smarter, but an email list serv would have been the best bet to get that strong user base.

Once you have the user base, then publish the mobile app. I just don't recommend starting out with it until you've proven the concept and have loyal users

How many web applications do non-tech people use? I suspect it's also less than 10. Probably closer to one (Gmail). I don't see why getting them to use your web app is any easier than getting them to use your mobile app.

> If I find a quality app

That's a big If. You're also under the assumption that the App is able to do something that cannot be done via a website. The sell is that the App offers a better experience to the user, often (e.g. reddit) that's only because they've nerfed the mobile website intentionally. The truth however is that an installed App gives them more opportunities to monetize you.

Whenever someone says to me "you should check out <snapgrambookspacenow>" I visit the website, if that website doesn't work on the desktop or directs me to download an App then I close the tab and forget it ever existed.

> The major advantage of the App Store (primarily for consumer apps though) is the fact that there's some kind of filtering at place.

Hah. Maybe when nearly every app doesn't have a dozen trackers pinging off before the app has even opened you can talk.

Right because web apps don't have dozen trackers on them.

I always read the first dozen or so reviews of an app before I install it. It takes only a few seconds to skim. There is a lot of value in that.

Kind of same for me. Before I switched to Android, on the iOS app store I was actually able to find interesting apps, but I guess I just had too much time to spend sifting through them lol. To me, the Google Play store is less organized. Now I seem to find apps more from others' recommendations.

How about start with a product that people would actually use? If it is a website, fine. If it is a mobile app, fine. Don't sacrifice business opportunities because of early tech choices.

To me, you have to have a really damn good reason to start with an app. Distribution is hell.

how so? click a button and download to device? now you have an icon that the user sees every day. How is the web better than that?

Are you saying you download more apps than you visit websites? I highly doubt it.

People only go to websites on mobile for one off things or they absolutely lack storage on their phone. Anything visited with any regularity demands an app instal, because bookmarks are clumsy on mobile and typing into the address bar on iOS at least has been frustrating since version 1 because of the lack of any spellcheck, even if you are very obviously typing a complete sentence or phrase, and how terrible moving the cursor can be with a touchscreen.

Not a lot of people know you can add a website to the homescreen, either.

there's a big difference between typing a URL into a web browser & downloading an app:

1. go to the app store 2. search for it and find the right one 3. make sure you have the right amount of space for it 4. download it 5. place it on the 5th screen of your phone 6. never see it

You can invent the same convuted path for everything. Ok, for the PWA: go to Google. Find the pwa you want. Make sure you did not exceed your data cap. Figure out how to add it to home screen. It will end app at screen #5 and will never be found again.

How is that easier? What if you want to sell your app?

Real scenario for native apps is much more simple: you get either the link to the app or the banner. Tap on it, tap get, tap open. That’s it.

Btw, there is an app offloading, which provides you space by getting rid of the app you don’t really use. And for the missing space argument to be even a little bit valid you must talk about huge apps, which probsbly means they provide functionality you could not replicate with a PWA.

Show someone with an iPhone how to turn any website into a web app with one button in safari and their mind will be blown, yet the feature predates the App Store. No one knows how to do this outside of tech circles. Apple had it as a stop gap before the coming App Store, then promptly never mentioned again to cement the store as the only way to get additional utility from the phone; web apps are a vestigial feature.

Challenges with apps:

1. Creating a good experience on Android and iOS

2. Getting approval to publish to the major app stores

3. Advertising to users so they actually download the app

4. Reminding users to use the app.

> now you have an icon that the user sees every day.

My home screen only has room for 24 items. Those items haven't changed much in a few years. I may have traded one app for another (e.g. Pocket Casts replaced Apple Podcasts), but that's about it. I don't look at the random Uniqlo app I downloaded to save 10% that one time every day. I don't look at the various restaurant apps every day.

I do have bookmarks in Chrome that I visit frequently. Why? I'm in Chrome for a significant part of the day.

You have those same challenges or worse with a website, except number 2. And for the vast majority of apps, that’s no problem. I’ve submitted 100+ apps to the App Store and never had an app that didn’t pretty easily get approved.

And I think most users have more than one screen of apps at this point. And are more likely to see those icons than a given browser bookmark.

I'm gonna play devils advocate here and say web apps aren't as pleasant to use on mobile phones as mobile apps and that's why mobile apps are still a thing.

I'm especially looking at things like touch gestures and native integrations (launcher icons, notifications, "share with" providers, maybe even a home screen widget). I know PWAs are a thing but afaik they don't solve all of the above problems or don't solve them well (for example there are still inconsistencies as to what is a PWA and how PWAs are handled across Firefox and Chrome).

Try https://usebx.com/app - we built it 2 years ago. It's a PWA, and our users find it better than many native alternatives. I think it's all in the execution really.

That's a cool app but it still launches inside a browser window on 3/4 tested browsers (Firefox on Android/Linux, Chromium on Linux; does launch as a standalone app on Android Chromium).

Also (but this is more of a niche problem with more technical users) I wouldn't have noticed it's a PWA if you hadn't told me since I block service workers by default. I would, however, have noticed a banner taking me to an app store of your choice.

Still, the app itself is cool though. Maybe you can use touch gestures on the web after all ;)

Beautiful app. Do you have a blog post about the development details like tech stack, etc?

Thank you :)

Not currently, but it's not a bad idea to write something. As soon as I get a spare few hours, I will put something together. Feel free to drop me an email (see profile) and I'll notify you when it's done.

This is so responsive! What stack do you use?

On the backend it's an async version of PHP (much like nodejs), and on the front end, vanilla js. We didn't use react or Vue as they can be rather restrictive (especially if you want animations), so we wrote our own little framework. Same with our CSS framework - rolled our own, because we wanted it to be really flexible across all devices.

Impressive, great work!

Most of my mobile users(64 %) has the page added to the home screen, and it's awesome.

There is however one problem in all things web: Apple.

The amount of bugs in Safari on iOS is incredible and the developers should be ashamed of their work. Safari is literally worse than IE6 was.

But sure, users can just switch to Chrome.. Oh, wait, we're talking iOS here, so there is only one engine, and that engine is broken. So the only option is to get a decent phone, aka not an iPhone.

Like, how freaking hard can it be to make a browser that doesn't need to be restarted just to make touch or keyboard working again?

Is there a way to know when a site is loaded from the home screen instead of the browser?

Yes, window.navigator.standalone

As the previous guy said you can check for standalone on iOS. You can also add an parameter to the URL so it opens eg. /?homescreen or something like that.

Apple has a financial incentive for making the browser as bad as possible for anything complicated. After all, they don’t get a 30% cut from browser apps.

Actually, they do get a cut from "browser apps" in the sense that they are paid by a big search/surveillance-based ad company to make their search engine default in the browser.

And again, they have a huge financial incentive to make the browser as good as possible because it makes the hardware as useful and wonderful as possible, which are currently the two most important factors for the most successful product that our planet has seen yet.

I'm kind of surprised by these 'browser is so bad' comments. Anyone here tried to make a browser before? Anyone make one less buggy than Safari?

>I'm kind of surprised by these 'browser is so bad' comments. Anyone here tried to make a browser before? Anyone make one less buggy than Safari?

So? By that argument almost nothing can be compared or criticized or discussed. Also, people aren't saying that making a browser is easy or that they can do it better or that Safari didn't do it well. They're saying that Safari did it badly compared to other existing mobile browsers. You don't need to build something to be able to compare it against something else.

>I'm kind of surprised by these 'browser is so bad' comments. Anyone here tried to make a browser before? Anyone make one less buggy than Safari?

not personally, but firefox and chrome both exist and are much better than Safari. Are you trying to argue that it's not possible for Safari to be better than it currently is?

No. I’m plainly stating, “really, you guys think Safari is horribly buggy or bad? I think it is really, really good.”

> Safari is literally worse than IE6 was.

You're joking, right?

I feel like this is coming from a good place. They're trying to help the majority of devs choose the right platform for their project to be the most successful.

As many before me have commented, the choice of platform depends on your situation, the technical limitations that come with it, and what kind of experience you want to deliver.

You have to use what makes sense and delivers what you need to satisfy your users. End of story.

Silver bullet solutions make a good headline, but they rarely address all possible concerns of a specific problem. Especially not with a broad audience.

IMO it's a misleading premise to begin with "Start with a website, not a mobile". One should try to stay true to target audience. If your product justifies "Mobile first" approach, mobile app is what one should start with.

Few years ago, I started with a website prototype with an end goal of porting UI elements and business logic to mobile. With some exception of reusing business logic, most UI elements required redo.

My main issue with app development, as I've been saying for close to a decade now, is that with _any_ app dev, you're automatically at the whim of whoever created the platform. Google, Apple, or Microsoft. We say a lot of things about how Google is trying to "take over" the web, but as long as there are Firefox users, that's unlikely to happen anytime soon. No one owns the web, that's why it's still the best place for a startup.

Especially Apple. Their iOS app development team uses the App Store as a proving ground for ideas that work well, then re-implement the features most users love in a crappy/cartoon-style way and attempt to out-promote the original idea.

Examples:

  Todos => Reminders,
  Flipboard => News,
  EasyMeasure => Measure,
  Skype => FaceTime

I don’t disagree with the main thrust of your position, but those examples for Apple are really poor ones, imo. Except for Measure, all of those are obviously features or default apps that a platform is going to offer. The fact that Apple didn’t have a reminders app at launch and then they added one later hardly means they ripped off “Todos”, which I’ve never even heard of. And Skype isn’t primarily an iOS app at all, which makes it a poor example of the phenomenon you’re describing. Also, Apple doesn’t make any money off those apps, so I’m really, really skeptical that they’re scanning the App Store for successful apps they can rip off.

> Web products get internationalization for free, which is important for a small, cash-strapped team.

I like the article but this statement is not true. Internationalization is quite a big topic and there is no way to automatically make a site internationalized for "free".

Don't forget that clients always get the lastest updated version of the app. I mean, I remember that everytime you decide to change something on the API, things go really bad because of the old versions and people just give you negative feedback because they didn't update the app. You have to be extremely cautious.

There are some very strange arguments in this article.

One claimed advantage of a web app is: "An App Store developer license costs $99/year, whereas a Namecheap domain is less than $10/month". Now, correct me if I'm wrong, but surely a year has 12 months, making it actually more expensive? I guess it's an advantage if you expect your business to fail within 10 months?

Also: "Browsers automatically translate text, so web developers don’t need to do much extra work to reach foreign users."

...

What? Presumably the author speaks no languages other than English. Developers of native applications could easily run their UI strings through machine translation if they wanted to - but nobody wants that because the results are at best unprofessional and usually nonsensical gibberish.

Typo. Namecheap domains are less than $10/year.

Surely it meant to say, $10/year for domain registration? Either way, this is still nonsense. Besides App Store developer license can be reused for other apps and comes with other additional benefits.

Depends on your applications use cases and what is needed to make it function. No matter what you're doing you should look for the quickest and cheapest option as long as it fulfills your needs. If I'm building pro grade video editing software I'm likely going to pick desktop technologies, if I'm building something that requires location or a camera I'm going to pick native apps. If the service is just reading data from a server and maybe writing a little with not too much interaction probably a mobile website or a PWA.

Not new advice, just pick the right tool for the job, not what's flashy.

There is one side of me that wishes more things were Mobile Apps. On the other hand, it's hard to argue that worse is better here and that any founder should focus on Web unless it's not possible to do it on the web (yet).

I'd much prefer a native app (or actually feels native; Well done RN is good enough for me), but the cost of creating and maintaining one is often too high when you don't know what will work.

As a web developer (react, etc.) I kind of wish I could have more time to work on native apps instead.

Is it a matter of UI? I find it interesting that the "web app with native UI" model never took on. There were a few tries (e.g. XUL), and now Expo seems to be a version of that for RN (though focused on testing), yet for various - seemingly circumstantial - reasons it never became a fully fledged alternative to browser apps and installed apps.

Because it solves no problems. Unless you consider “we want web tech everyhere” to be a valid problem. “Cross-platform” development may look like a problem worth solving, but the dolution is a lie. If you use native UI you will have to deal with a native tech to a degree. Is (cost + pain)/benefits worth it—that’s up to you. But seeing how RN stuggled with the thing that’s trivial in native—navigation stack—I’d stay away. Also these bridged technologies will always lag behind and can be really painful after significant updates to UIKit.

> An App Store developer license costs $99/year, whereas a Namecheap domain is less than $10/month.

I think you mean $10/year.

Web hosting will cost more than $99/year so not all the costs of the web application are factored into this article.

Sure, but there's a bit of apples and oranges here - for example, any modern internet-connected app would also need a server backend.

But once we talking about that money, $100 a year is peanuts.

True. It's also worth noting that a $5/mo VPS can get you surprisingly far with respect to well-designed websites and services.

You shouldn't run a production environment on a single $5/mo VPS.

It depends on your speed, concurrent user capacity, and uptime requirements (among other things). I run two production servers off two separate $5/mo VPSs, one getting a few thousand hits a day. No issues whatsoever.

What happens if something catastrophic occurs?

What do you mean by catastrophic? It's safe from the obvious things:

* DB corruption / destruction: The database is backed up.

* FS corruption / destruction: The production code is backed up with version control (obviously) and can be restored at the drop of a hat.

* The server just randomly falls off the internet: It would take about 2 hours to get a new server up to the same point.

What kind of problems could arise, that wouldn't equally affect any other kind of server provision?

How frequently is that database backed up? How many hours of work are users going to lose? And you're going to wake up at 3:00am and take 2 hours to get a new server online?

This is all fine for a service with a few thousand users that's not mission critical. But I assume the article is about building apps to be a real business not just a hobby.

I run a few hobby sites under the same scenario as you but all my work sites have multiple servers to fail over to and replicated databases.

That all sounds very reasonable. Ours is a news website which publishes semi-frequently; so we'd probably lose about 5-10 articles on average. It's not just a hobby - it brings in more than a minimum job's worth per month at the moment and is growing fairly quickly, since we operate in a growing niche.

Like I said - it's all about your requirements, and for us $5/mo is just fine.

Also I just installed Daylio which is a cool little diary application. They have no web app, it's mobile app only. It backs up your data to Google Drive. It has no server component, no signup, etc.

Admittedly there are limitations of not having a server backend but for a MVP that's another way to go.

And apparently no way to update from my desktop then.

Exactly. No API either. So obviously it's not without it's downsides.

A native app with no server backend is the same as a web application with no server backend. In that case you really just need to put those static web files into an S3 bucket and route a domain to it. The cost is very low for this.

Athough localstorage exists, I'd be very hesitant to store anything important/permanent there. So your no server web app will be pretty limited.

Daylio is good! I did it for about 6 months but I've tailed off recently.

Simple apps can definitely get away without a remote backend, and for things like single-player games they can get away without any web presence at all. So yeah, there's definitely no one true solution.

If $100 per year for the Apple Developer program is too much, is whatever you’d be making an app for even worth it? Not saying app-first is or isn’t the answer, just questioning a criticism of an expense that’s comes out to $8.25 per month.

That expense isn't high, but you also have to own (unless something has changed in recent years) a Apple device to develop on and an Apple device to test on. I own a Linux laptop and an Android phone. If I wanted to write an iOS app I think I would be needing about $2k in devices.

That's my feelings, although I've seen so many successful app-only startups that I feel I'm missing something that these startups grasp.

E.g. in the UK Monzo, an app-only bank, won't allow you to access your statements, do transfers, or pretty much do anything to manage your account on the web, and is one of the most popular challenger banks.

Most of the points in the article are for the benefit of the developer (easier updating, lower dev cost, fewer platforms) and not for the benefit of the end user. If the goal is to make a product better for yourself then make a web application. If your goal is to make a product users will like then make mobile applications.

The article talks about how users don't install apps but signing up for a website (enter email, invent new password, wait for password vertification email, click link) is a far larger barrier to entry than just finding the app in the store and installing it.

Just because everyone and their mother adds a signup flow just to be able to use their web app doesn't mean it has to be done that way. You can add the signup flow after you've already demonstrated the value to the user and when it would provide additional value, like saving things to your account.

An anonymous user on the web is equivalent to some anonymous user downloading your app from the app store.

It's true. And the article's example actually doesn't have a signup. But often demonstrating value to the user will require multiple accesses and you don't want them to lose the data they've already entered. It's probably better to force a signup then potentially lose everything the end user has done in your app.

For a mobile app, you can store data locally or on a server keyed to the user without any potential for loss and without a signup.

There are users who prefer to use a desktop web application than a mobile app. Many types of task are better on a large screen with a mouse and keyboard.

It depends on the type of app and target audience.

> desktop web application

Web applications are significantly better on the desktop than on mobile so I completely agree. In fact, if you want to support mobile and desktop I'd recommend apps for mobile and a web application for desktop.

But if you expect that a lot of users will be using mobile, the experience of accessing and using an app from mobile isn't that great.

this article is great advice for solo or small group of devs to test the concept first. a startup with funding or any company with tons of rev has plenty of money to plunge into the proper mobile dev team and also do user research to see what they'd want to have in the app to use it.

> successful app-only startups

survivor-ship bias. You don't see the thousands of app only startups that fail every year.

In the case of mobile banking, your smartphone acts as a second factor of authentication. Older style UK banks require you to have a separate authentication card reader for authentication of transfers, and use the card reader to generate a code to input to the website. Making it app only gets you the same security with more convenience and less cost.

This has always been our philosophy with AutoTempest.com. We do pretty regularly get requests to build an app though, although none of the people asking for it have had an answer to the question, "What are you looking for from an app version that the website doesn't currently offer?" So, at least so far I've decided it makes more sense to focus resources on the website. I do occasionally wonder if I'm passing up on a separate avenue for growth though. I also just wonder why people ask for an app, when they don't have a sense of what they want from it. There's obviously a sense that an app version would be better somehow, but it would be nice to understand what sort of betterness people are anticipating. (As someone who will almost always choose to use a mobile website over installing another app, I really don't have an intuitive understanding of the desire for an app.)

As an aside — I would like to say that I really enjoy your website and the service! It has made shopping for a decent deal in my area a breeze.

I'm curious about what kind of statistics you gather? It would be neat to track price movements across regions, and to see how the market trends for specific body styles and other characteristics during different time periods.

Thanks, glad it's helpful! So far we only keep statistics on searches, not on listings. We do plan to start that eventually, but so far there have always been more important things to work on. (Important in terms of having more of an impact on the core search experience.)

Could it be as simple as an icon on the home screen and separate entry in app switcher?

We do have one of those notifications that shows people how to add the site to their home screen, so hopefully that catches most of that use case.

I read recently that only 0.01% of android apps recoup their development costs. I only understand android apps in terms of supporting other existing products and services in that context

https://www.gartner.com/newsroom/id/2648515

Every single app I install immediately starts showing me upsell notifications. So the first thing I always have to do after installing an app is go to the Android app manager and disable its notifications.

Websites don't have this issue. This is just one of many reasons why I'd rather go to your website than install your app.

(disclosure: biased iOS dev)

If your app contains significant user interaction and input, I strongly prefer a native app. If I'm just consuming text, then I prefer a website.

Incidental to the main topic of the article (which I hope I don't distract from), but:

> Browsers automatically translate text,

Just Chrome not "browsers", no?

> According to research from the Chrome team at Google and others, people spend more time on mobile apps than they do on mobile websites. [...] 90% [of app time] is spent in a person’s five top apps.

To me this just sounds like people are drawing incorrect conclusions from the statistics? It doesn't seem like people spend more time on apps necessarily but that they download the apps for the things they use the most.

I imagine most people are less likely to try something if it's only available as an app. Having to download something just increases the barrier to entry like a very long page load.

so one assumption that could be made is that mobileapp use takes longer to accomplish anything compared to using web

Antidotal (likely an outlier) but I’ve slowly deleted social apps (Facebook, Twitter, insta) in favor of using their mobile sites. My reasoning is three fold: mobile sites are pretty good, it’s harder for companies to mine my data in a browser, and honestly storage on my 16gig iPhone is far too valuable for a 300mb app!! I’m 18 (aka young and woke) btw.

Side point: I still need Snap’s app but I’ve considered quitting altogether bc it’s barely usable on my 6S. If battery is < 10% my phone will die if I open SC

> it’s harder for companies to mine my data in a browser

I agree with this but I'm curious: What protection measures do you take to prevent companies from mining your data in browsers?

Android actually covers many of the "pros" listed here. It's cheap and fast to deploy. It's also easier to SEO an Android app than a website, from my experience. Mobile searches will turn up your app. Customer acquisition cost can also be very low.

I'll add one other reason to use apps - it's best suited for things that happen as a kind of habit. For example a meditation app or a to do list is going to be more successful as an app than site.

My simple take on this is "holy shit yes!" Apps are the new desktop client, as far as I'm concerned, with their "innovative" UIs, inscrutable error modes, inevitable lack of support for still usable phones, and stupidly invasive permission requests. No web site ever needed access to my contacts, and every time an app asks for it I look for a "fuck off and ban-kill this app" button.

Easier said than done. This is why when everyone has switched to React Native we have stuck with PhoneGap. Might we be ridiculed? Maybe. But we can easily make apps like this:

https://coinect.app

Here is my logic!

1. Invites have a URL.

2. The person doesn’t have the app yet, so they can get taken to the Website or the app store.

3. If Website, then if invite token is unique we can automatically confirm their phone number or email, and get everyone who has them in their contacts to follow them, so they instantly get friends and socially generated content they can interact with.

4. And then redirect to the app store. Then when the app is downloaded, it can open a BrowserTab inside and get a signed NewSessionID for the app. (Note: this ain’t oauth folks, our server uses a secret to sign and verify its own issued payloads that are relayed through clients like this.)

5. The result: zero passwords needed, tons of steps skipped. But then you think, what do you need the app for? Android has Chrome Notifications. So you only need it for Contacts. So you place buttons to download app in appropriate places.

6. If you’re gonna make a website anyway, then why not reuse it inside WebViews to not duplicate work?

7. But it’s not easy to make a website working across all browsers and devices and take advantage of touch and cameras on phones, and work differently on mobile phones (with sms://...) and QR codes and so on. There has to be a platform that does everything for you so you can just focus on building your app. Same as Mac and Windows handled GUI for all apps so you could focus on building your app. We need a Web OS for apps.

8. So basically, after 7 years of work, you arrive at this:

https://qbix.com/platform

Or better yet, why not make your website the app? Flutter.io comes to mind. This will save you considerable time because you won't have to maintain as much native code on android or ios since your app is essentially a native browser window. You will only have one small codebase to maintain and it is cross platform.

What is funny, is this is the opposite I heard on the podcast "The Pitch" when all of the investors were complaining that the developer built a web app first. I can't remember the product, but he justified it well, eventually they agreed, but wanted to see an app before investing.

Reminds me of the time when WIRED magazine back in 2010 declared the web 'dead.' [0]

Kind of amusing to note how wrong they were.

[0] https://www.wired.com/2010/08/ff-webrip/

This reminds me of the time back in 1997 when WIRED offered suggestions to save a dying Apple [0]. First suggestion? "Admit it. You're out of the hardware game."

[0] https://www.wired.com/1997/06/apple-3/

I had similar idea for my 1st game. I don't like the idea of downloading native app for everything.

I bought a (.)app domain to host my game. I never build website before, although I have decent understanding of HTML and CSS, so people suggested me to start with a static website first in HTML+CSS. Result was a terrible website www.chessxplode.com

Next was to learn JS, React, HTML5/WebGL, cloud computing and more to build the game. Instead I decided to focus on what I know well, game programming in native iOS. Once native game was ready, focus shifted to grow the game.

Gaming is one of the use cases where native wins, specially since WebGL 2.0 isn't widespread and it only exposes a subset of OpenGL ES 3.0, let alone GL ES 3.1, 3.2, Metal or Vulkan.

However if the game is a Flash like one, then WebGL/WebAssembly might just be enough.

Macro stats about install/uninstall rates of mobile apps are meaningless distractions when you're building a startup. Focus on what your audience needs and build that, whether it's a mobile app or a website.

I built a funny videos android app and website, android users discovered the app more easily and were more engaged with the app because they wanted to share videos from my app to whatsapp.

My more recent app is a web app because my audience are people doing work on their computer.

Audience guides the approach.

I built a crypto wallet / portfolio tracker that is much more useful as a mobile app. I can imagine how expanding it to a website could make sense down the road, but the utility is in quick access to transfer funds and check your portfolio which would be a pain on mobile web.

I used Expo (platform for React Native) which made it arguably easier than deploying a website. I also get instant updates too without having to resubmit to the play/app store, and its cross platform.

Start with where the money is, which is usually iOS, which doesn't support new web technologies. Guess we're stuck with apps after all.

Lets say that there are 2 products, both of them offering similar functionality which is useful to a user. One of them only offers a progressive webapp and the other a native app. I can't think of a single example of where a product offering the former has won over the latter.

Does anyone know of any products which have succeeded despite not having a mobile app ( while their competitors did)?

Always start with a website and only build an app when the limitations of a website hinder growth and/or experience.

If you only want an app to be listed in the App Store, send push notifications or get a tile on users' home screens then use something like https://appwraps.co/

Would it be too much to take this one step further and just say, you don’t even need a website? https://link.medium.com/LgwqnAIxgT

Your business idea doesn’t have to be rooted in anything technological. You can still start something offline, even in 2019!

Can’t do in app purchases with websites. App Store does a lot to help your app a chance to get discovered without SEO. Google will likely list thousands in front of you where as App Store will get you up there unless you are in a most popular category. Of course, lots of ideas don’t fit apps such as company websites or bios

This has not been my impression at all. With Google you have a chance to do SEO and improve your ranking and discoverability. With the App Store your only chance of discovery is their crappy search box where you compete with 100s of spam apps that somehow made it through the “rigorous” app store approval process.

A handful of social media platforms have launched that I've wholly missed because they didn't bother to have a web version. As a Windows Mobile user, I couldn't install your app if I wanted to, and I spend a lot more time at my desk than staring at my phone.

As a Windows Mobile user, you're too small and niche of a market to care about.

> "whereas a Namecheap domain is less than $10/month."

Small typo there, I think you mean $10 a year.

And you have to add: Don't build a heavy front-end application. Build a small server first and return HTML. Once you prove your product then use something like Jasonette to serialize your model into an app language if you want to, or stay with a website, why not?

> So, we built the web’s first video meme generator. Within weeks, Kapwing had hundreds of meme creators using the website daily thanks to the magic of SEO.

I want to learn this magic too! OP can you give us a brief overview about your SEO strategy?

Of course I'm not interested in downloading a friggin app for Dominos pizza

And I also don't believe I'll ever find a webapp that can favorably compare with my notes/calendar/email/other utility app of choice.

The most important factor is the cost. Developing and maintaining a mobile app is expensive, and often there is no justification for spending that money, especially given that most users expect apps to be free.

It's a great point about web first. As a side note, there's really no point to a college degree anymore for most people. You don't need a Stanford degree to make a leisure product.

When 5G is pervasive and cheap, "instant apps" will undoubtedly become a thing, which will reinvigorate apps. The App/Play stores cause too much friction now. That will change.

"Browsers automatically translate text". I was not aware of non-Chrome major browsers having this feature, what have I missed ? Are they in fact targeting Chrome first ?

Using a separate phone for apps and using another phone for calls/texts can help privacy if you still love to install games/apps

Or just use iOS, because unless you’ve heard about an exploit that I haven’t, it’s impossible for any app to read (or write) your texts or calls. Or anything at all outside its sandbox unless you explicitly give it permission. This includes location, photos, contacts, microphone, etc.

In fact, were I a malicious developer instead of the benevolent one that I am, I don’t think I’d touch iOS unless I could find a security hole or I was going to trick users into giving me permissions, because I don’t think there’s much attack surface otherwise. Open to correction though!

Domain is $10 but it’s for a single domain. For every website you will buy one. In addition, play store asks one-time $25 fee. You can submit as many apps to app stores as you want.

Generally, you have to do more things during web development such as SEO, cross-browser testing, meta tags. In addition, you need to deal with responsive design. Mobile apps are easier easier to develop because apps display one thing at once while web sites and apps display a few components in one screen.

On technical aspect, it definitely web first. Before a native apps, it need web app --> web services --> native apps.

Do the PWAs have the option to be shared to when added to home-screen? Like, can you share to them from other apps?

Not as of yet, at least not with any PWA I've used.

On UWP and ChromeOS you can do everything if the PWA is packaged via the store, as they get access to native APIs.

https://developers.google.com/web/updates/2018/12/web-share-...

Most companies hire me to build a mobile app because they think App/Play-Store are the places to be, marketing wise.

The author seems confused about backend vs client app.

"An API Kapwing depended on broke unexpectedly, and we shipped changes the same day to get our service back online. Since we could deploy immediately, the unexpected outage was only a small hiccup. If Kapwing was an app, it would have taken a week for the new version to get approved, go live, and rollout to users through app updates—that would have been a huge loss."

another reason to build for the web: https://android.jlelse.eu/google-just-terminated-our-start-u...

"Namecheap domain is less than $10/month" I believe the author meant $10/year.

start with a shellscript, then a REST API (+REST aware CLI), then a website, then a mobile app.

And sometimes even start with a mailing list (especially anything marketplace related)

I am moving towards "Windows app first"; Windows still has dominant market share among Desktop users. You might impress investors and developers with Mac and Linux support, but far too many people are developing for platforms where the users are not and wondering why their stuff isn't catching on.

I've been trying to find time to give Avalonia (a cross-platform desktop GUI kit for .NET Core) a try.

https://ptupitsyn.github.io/Cross-Platform-DotNet-UI/

Looks interesting. I've been curious about the state of .NET Core for cross-platform applications. Rant: Why does every cross-platform framework target GTK on Linux rather than Qt? [1]

[1] https://github.com/AvaloniaUI/Avalonia

Windows OS has a dominant market share yes, but do those people download and install a lot of desktop applications? Especially on a platform like Windows, where a browser can do almost anything a native app can do?

The counter-argument is that it is that everyone has a web site and if you want to make something that is really different and better you need to do something different.

For instance I use em client as a mail program and fastmail for the back end. It's a lot like MS Outlook except that it actually works. I prefer it to GMail.

With the rush to Electron we see developers leaving so much performance on the table, particularly with the rapidly increasing core counts.

There's also an argument that the "web" is no longer an open platform but is now becoming the "Google Platform". For instance you cannot trust conventional web sites from a privacy viewpoint now that most of them have numerous third-party trackers, download content from multiple CDNs, etc. This deviance is not so normalized for desktop apps.

Meh... I'd lean towards web app first, wrapping in cordova for ios/android and electron for desktop at first... adding in more native features as time moves forward, probably react native for {platform} as a second stage.

What I am aiming for is an application that never locks up and never makes you wait without clear indication if it is making progress or not. I could care less about many "Native Features" but latency is king. I am sick and tired of looking at spinning beach balls and such.

If your data backend is over the internet, it doesn't matter if it's a "native app" or "pwa". You can handle appropriate state in most cases. If you're good at it, you can do it at the most atomic level, and should be relatively quick. If you can shard your api/data systems by locality, even better.

It's the 95% latency that kills you.

I have seen so many people make an app that is blazing fast at the office (either on a LAN or w/ a fast connection to the cloud) that ends up horrible for many end users. Sometimes the people who made the app don't even know.

start with an app because Apple is more restrictive than the web and ideally you are going to want both an app and a web app, you don't want to fall into a situation where your popular web app is not allowed as an iOS app.

Lots of regular users have issues with limited space with which to download new apps. Of installing a new app means having to get rid of one I use, then your all just doesn't get installed. Stop building apps that could work as websites. It's not 2012 anymore.

"Start with a Website, Not a Mobile App"

It sounds pretty obvious to me, low barrier to have something working = early feedback. Given the number of upvotes, I'm impressed why this is such a taboo. I always create a website first.

Another pitfall on the Apple Store at least is that you as developer or agency creating apps for customers have to create an account for each entitiy (customer, company). Another $99, another excruciating approval process.

Thank you! Finally someone says that here on HN \o/

What about if you're making a video game?

Wouldn't you have the best of both worlds with Progressive Web Apps? I thought that development strategy is supposed to solve this very problem.

i agree, its way easier to purchase a domain and launch a site!

case by case. you cannot start uber via a website.

Go PWA

finally, someone said it.

I'd like to venture another guess even further that this is possibly the furthest venture anyone can guess.

I'd venture you'd be wrong. I'd not venture to guess how deep one could venture.

I agree and I would go as far as distancing myself further to guess that you just ventured one guess deeper.

Please don't.

We detached this subthread from https://news.ycombinator.com/item?id=18828531 and marked it off-topic.

Mobile web apps are the future, thanks to WebAssembly.

This is wishful thinking with no logical argument behind it. WebAssembly replaces JavaScript but the rest of the experience (a web browser) is exactly the same. What would compel users to care?

> This is wishful thinking with no logical argument behind it.

1. A web app developed using WebAssembly offers performance indistinguishable from that of a native app.

2. Since the web app offers native performance, there's practically no need for a native app to exist in the first place. Why should developers waste their time developing for multiple platforms when a single mobile app works flawlessly?

3. This is great for users because the app is highly performant, universally accessible, and doesn't require installation.

> A web app developed using WebAssembly offers performance indistinguishable from that of a native app.

If the performance of your app is limited by the DOM, this will make no difference. Most applications are not doing heavily calculations, they're animating an non-native HTML UI so WebAssembly will make no difference. Raw calculation performance is not the biggest issue for mobile apps.

> Why should developers waste their time developing for multiple platforms when a single mobile app works flawlessly?

Nothing about WebAssembly will make web apps work "more flawlessly" than web apps work right now.

Since WebAssembly doesn't change the user experience of acquiring and using applications it will make no difference, from today, on adoption on web applications.

I'd go so far as to say that users actually don't like going into Mobile Safari run mobile applications and WebAssembly isn't going to change that.

Javascript is already fast enough for native apps as can be seen with things like cordova. Even with WebAssemply the layout and ui draw will still need a browser engine which is where the real difference between native and mobile is.

1. I'll believe it when I see it. 2. The platforms have different assumptions and affordances, and trying to homogenize the experience by splitting the difference (like you seem to suggest) will leave users in the uncanny valley. 3. No, it's great for web developers who want to claim they do "native development" without actually learning how to do native development.

WebAssebly offers zero of the things native SDK offers. I wish people making these claims took at least a cursory look what’s in UIKit alone.

I feel like all we need is "one Ring to rule them all" - one app to run on all the mobile and desktop OSes and in web browsers and I can see no reason why do we need different kinds of apps for these (there hardly even is such a thing as a web site today actually - all the websites, this included, are apps, just styled and implemented differently and optimized for bigger screens). Perhaps Flutter (or something else to come) is going to let us achieve this.

Reply on news.ycombinator.com