Hacker News Reader

> ploddingly slow with clunky non-standard UIs—a degraded, lackluster experience

That sounds like the arguments I used to hear from desktop devs back when web apps start to appear.

Sure web mobile apps kind of suck right now, but theres hardely any frameworks, established processes or documentation on how to build them properly. It's still in alpha stages.

When building mobile apps has matured and if they still have the performance and UI issues, then I might agree. But right now thats like saying a kid won't have a good job because of his grade 5 scores.

That sounds like the arguments I used to hear from desktop devs back when web apps start to appear.

It's also like the arguments made against Java 10 years ago. :) I agree with you - it's not yet clear whether web technologies will significantly replace native for mobile or not. Even if they eventually end up doing so, that appears far enough out that I can't imagine making any decisions based on that right now.

I think the current popularity of web apps over desktop apps is driven by two different motivating factors: social and centralized storage.

The nature of the Web makes it well suited to building social apps on top of it. It would be hard to build social apps on an application layer that didn’t transcend physical platforms, because then there would be platform-induced delineations between the social groups that could congregate.

But the other motivating factor is that web apps by definition are a centralized container for the data you put into them. You use Freshbooks and you can access your invoices from anywhere. I think this will become far less important as cloud-storage-on-the-desktop services like iCloud mature and largely eliminate this advantage of web apps. Given this, don’t you think think we’ll actually start seeing a resurgence of non-social desktop apps that would previously have been built on the Web for centralized storage reasons?

Web apps beat native apps because there's no install step, no upgrade step and hyperlinks provide the ultimate way for different apps to communicate with each other.

Agreed, but I would generalize it a bit more and say that web apps reduced the barrier to consumption considerably over desktop applications.

In most cases, mobile apps are close enough to web apps in regards to barrier of consumption to make the difference moot, and when you add the native advantage, the result is a net positive consumer experience for mobile apps.

install once is often easier than navigating to a URL on a phone each time. Many users like apps due to one click launch

When you find a web app you like, you can put it front and center on your Home screen. Just open the web app on your iPhone or iPod touch, tap the plus sign, and then tap “Add to Home screen.” A Web Clip will be added to your Home screen automatically for easy, one-tap access.

http://www.apple.com/webapps/whatarewebapps.html

not sure about android, and 80% of users will not do this (like they do not use bookmarklets). Also many web apps require you to re-login after a certain period, which adds even more pain compared to apps.

"many web apps" do. But they don't have to. That's a choice the developer makes. That's super easy to fix, too.

Agreed. Like Google+ and gTalk etc? This is clearly the UX standard ahead yes?

Web apps are fine for lightweight, data-driven tasks. For tasks that require richer interactions they are still decidedly inferior. This may change as HTML standards and compliance evolve but it's not crazy to suggest that a well-written native app will always be one or two steps ahead.

The question is: besides games, how many apps really use richer interactions than what's already possible using web technologies?

All multimedia content creation apps. Think DJ apps, music creation, audio editing, video editing, 3d modeling, image editing and illustration, etc. Basically everything that makes computers creative tools instead of information management devices. I don't see web versions of any of these seriously challenging native any time soon. Hell, even spreadsheets and word processors are still better in native.

I don't understand why this has to be an either-or thing. I know it's easier for us to think in black and white terms but I don't see why web and native can't coexist for a long time.

Those are all desktop apps, though. Mobile apps - like the ones referred to by the article - are rarely for content creation.

Not yet, but these things are converging. There are already fledgling implementations of a lot of these tools available for iPads.

From your comment it becomes clear: At the moment "web" is primarily a data display platform. Not surprising, given its pedigree. Input techniques are quite limited... Especially responsive ones we've become used to on mobile. Sounds like a good project to work on..

There _is_ an install step to almost all web apps, and it generally involves having to fill out a boring form, clicking through from a confirmation email, and trying to remember a username and password every week for the rest of the app's useful life.

This is unique to applications that wish to store users' data in a central location.

I suspect we have different definitions of 'unique' and/or 'application', because this sounds to me like someone saying the need for tyres is unique to cars that have wheels.

Perhaps. Do you consider a Paint or Paint.NET clone to be an application? A calculator? A clock/weather app? A game where you throw birds at pigs? These are all samples of potentially native applications that do not require login and can be implemented as web pages.

How about services that already require you have an account? You have to log into a Twitter or Facebook or Linkedin client even if you install it as a native application.

I don't care that native apps require work to install. We've already accepted that as a premise. What I'm saying is that web apps have setup costs too, and therefore saying you don't have to install them is disingenuous.

If all you require to get through the day is an online paint clone, a calculator, a clock/weather app and a game where you throw birds at pigs, more power to you - I'm glad you've been able to shave a few seconds off your day.

"used to"?

I certainly haven't stopped making this argument for web vs. desktop apps, as it has yet to be solved. If it even can be solved (mainly the UI/UX issues; performance probably will eventually approach "good enough" for a given app.)

>But right now thats like saying a kid won't have a good job because of his grade 5 scores.

Conversely, would you say that a kid in grade 5 who has bad scores is going to have a great job when he grows up, just you wait and see? We can only judge the mobile web as a platform (if one can even say that it is one platform and not many) by what it offers on mobile handsets today, not by what we hope it will one day become. Also, we can't judge it by hoping that one day, all the mobile browsers will conform to the WC3's standards, because we know how good a job today's desktop browsers do at quickly becoming standards-compliant.

Here's the fact: The web is a computing paradigm that was not designed for mobile.

A mobile device is just another client. The web was designed to be agnostic in that regard, was it not?

no it was not. hence the need for "mobile" versions of websites and various hacks to make the web work on mobile devices.

How does that follow? That's saying that the mobile clients don't implement everything, not that http isn't agnostic.

that's not the web's fault... that's the fault of most developers doing the web wrong, refusing to separate content from presentation.

If mobile users only need service-type apps, the web can provide that. Programmers just have to write the screens.

The Web is a dumb terminal. You fill in a form, press XMIT or the equivalent, and get the next screen. Just because someone decided that the content needs to be jammed in a 900px table cell doesn't mean the web can't do mobile too.

"When you have a hammer everything looks like a nail"

The web is just another computing paradigm with its own strengths and weaknesses. It's not the be all and end all of mobile.

So? The web was not designed for video. It wasn't designed for dynamic interactivity either or applications at all, actually. Yet here we are.

That sounds like the arguments I used to hear from desktop devs back when web apps start to appear.

And it still holds true in many cases. Microsoft Outlook, in many ways an example of both great and horrible desktop software, still holds the email client majority: http://www.campaignmonitor.com/stats/email-clients/

Web apps are great when you get along with the designer's UI decisions. When you have a fast network connection (that's improved with specific web apps and offline modes, but it's by no means generally applied). When the servers are working well.

Mobile apps share the same advantages and disadvantages, but they're multiplied tremendously. I lose my mobile signal in most stores I go into, for example, and on an awful lot of secondary roads throughout the Northwest.

Agreed. And for some apps, where you need to work offline, native is great. But for apps that require a network connection to work, I don't see the point of a native app.

Outlook? The "majority" email client? Really? From your source:

"The fine print

The email client a person is using can only be detected if images are displayed. This can give an inflated weighting to email clients that display images by default, such as Outlook 2000 and the iPhone. It will also provide a lesser weighting to those that block images by default such as Gmail and Outlook 2007. Those email clients that aren't capable of displaying images, such as older Blackberry models and other mobile devices cannot be included in this study."

Outlook isn't nearly the majority email client, it's just the one most likely to not block img bugs.

Outlook isn't nearly the majority email client, it's just the one most likely to not block img bugs.

Do you have anything to support that claim?

> Sure web mobile apps kind of suck right now, but theres hardely any frameworks, established processes or documentation on how to build them properly.

Your argument assumes that building good web applications is merely a function of tools. It's not.

We have a performance-per-watt wall, plus a heat wall, plus a weight wall in mobile hardware.

There is a reason that Apple is pushing C-based development without as much as a garbage collector, it's not just the stubbornness of Jobs. It's because a dynamic language with state-of-the-art VMs struggles hard to do basic smooth animations on current-gen mobile hardware, and will continue to struggle, for at least the next five releases of Nitro.

You're going to need something like a battery revolution to build the mobile web app panacea that you want.

It is a function of tools. iPhone App have smooth animation because they run on Open GL accelerated graphics, yes. But so could Android apps. The difference is that the iOS SDK allows you to `pushViewController:animated` and that's it.

In the (pretty far I think, unfortunately) future, an SDK/Framework/Library à la jQuery will propose the same tool to run Web GL animations.

We need a UIKit/CoreAnimation for the web. The technology is not there yet, but it's going to be.

I realize you're referring to Swing Java apps, and I won't argue on their behalf for a second. However the SWT library is a dream of a native toolkit experience if you're writing a client app in Java. It's fast, does a good job of working across a wide variety of platforms (Win32, Cocoa, GTK, and I think Motif), and is extremely well documented.

Web apps don't have to be as "good" as native apps to succeed. They just have to be good enough. Remember that web mail exploded in popularity in the late 90s. Do you remember what the experience was like on Yahoo Mail in 1999? Full page refreshes on dial-up modems to see a single email. Yet people swarmed to Yahoo Mail, Hotmail, and the like.

The problem here is that these articles are very extremist. There are many native apps that can easily be done as web apps (cost and time effective), but not all of them. You can't choose one platform and just say it's the best choice overall.

One significant difference between Java apps and web apps is that, unlike Java, the platform vendor has huge incentives to keep improving your runtime (the browser).

Get used to these kinds of comments. There's a large contingent of people on HN who don't understand that higher level languages (say, Java) are not a complete replacement for C/C++. The same people are behind the "HTML5+Javascript ought to be enough for anybody" mentality.