Comment by anexprogrammer | Hacker News Reader

Comment by anexprogrammer | Hacker News Reader

Parent Full thread

> But if that means creating a unique UI and app flow for long-term user happiness, then the app company owes no respect to the "local conventions" of the OS.

This PoV has engendered more hatred in me towards software creators than probably any other. Not bad considering I spent 25 years creating software.

If I'm on Gnome/Android/Windows 10/OSX/Whatever, frankly I'd prefer everything follows GUI guidelines, standards guides and looks, acts and feels like a native. Trouble is everyone knows best.

iTunes "knew better" and made Windows look and feel like a Mac. Fine, except it made the program clunky and stand out for being different. Apple knew best with iMessage and decides it should invisibly merge sms and their own messaging. Result? A dozen times a day it would hang for 10 minutes or so trying to connect to data-unreachable imessage servers. Just send a damn SMS, which is what I wanted in the first place. Google came up with Hangouts that at first release (i've not revisited it) managed to leave out every essential of a messaging app. It was sort-of pretty, but very broken.

Google and Microsoft always know best and don't even respect their own standards, let alone anyone else's. Neither are any good at all at designing GUIs and frankly should just stop. Outsource to someone who can. Samsung should take note of this too.

I don't want 40 different unique UIs all optimising app flow for long term user happiness as that's anarchy with 50 (sic) different ways of doing things in those 40 apps. I'd prefer it felt a little dated to work on Android or whatever if it was then seamlessly perfect in gestures, GUI and function.

Everyone needs to just stop knowing best. 99.9 times out of 100 they don't.

I was really impressed by WhatsApp's approach. When I installed in on my iPhone, it looked just like the native iOS messaging app. When I helped my mother-in-law install it on her Android, it had a totally different look, much closer to Android's native messenger. Both of us pretty much instantly knew how to use it because we had been all along.

> Both of us pretty much instantly knew how to use it because we had been all along.

Exactly why going native is a really good thing - people can fully adopt your app almost immediately as they don't need to learn your different (knowing best) conventions and extensions.

The difference is Whatsapp isn't intended to be used outside of mobile, where it's atypical for the majority of users to have both an iOS and Android device. With the Google suite of applications however (most prominently Drive), the same user is expected to access the service across a variety of devices. A consistent design means they don't have to re-learn how to use it.

Totally fair point. In both cases, I think design follows the Rule of Least Surprise to create a consistent experience for users. It's about business aims, too. WhatsApp went after native messaging, so it made sense for them to make the transition as seamless as possible. Google Drive applications are operating on a completely different scale.

At the same time... Google Drive applications looked very familiar to people coming from the MS Office world. If you're trying to replace a competitor, mimicry isn't a bad place to start.

If you're using a mobile version of a desktop/tablet application, you're going to have to relearn it anyway because of the differences in size. I don't know many users who operate multiple mobile devices of the same size but different OS simultaneously (say, iphone and samsung galaxy), so I don't see why such devices should share UI designs.

Facebook Messenger is also extremely similar to iOS's built-in messaging app, which is nice. In fact, a fair amount of the design improves upon Messages' already pretty-solid design.

But their Android version also looks pretty much exactly like the iOS one, same for their web version.

Facebook is going down the consistency across platforms route and not the consistency with platform route (I'm not passing judgement on them, and in some ways I disagree with the article's main thesis).

The merging of iMessage and SMS is something that I've always thought Apple did really well. I don't care how the message is sent, as long as it is sent. Create a quick hierarchy and send in preferable order. That's exactly what it does.

I spend a lot of time without either signal or internet when in the countryside and this allows me to seamlessly communicate. The last thing I want is to either switch programs to carry on conversations, or manually choose what kind of message to send. Both would be a pain.

My main objectbbion to iMessage is that it only works with other apple products, so I end up using whatsapp with some people - where I need 2 apps to handle simple messaging and conversation histories get split up etc etc.

I did too until I changed work location from always great signal to marginal. Lots of phone apps don't handle this gracefully (I think most expect perfect data). IMessage could have popped a dialog at 10s (30s), but decided to wait forever instead. Then I started to dislike it. This is a good time ago now, they may have sorted this, I've been Android the last 5 years or so. :)

Agree on the messaging. I miss just having adium or pidgin connected to everything, now I'd need a dozen chat apps to stay in touch with everyone!

iMessage is "ok" now. I'm not sure the time duration, but I can usually long-press a message that isn't sending and tell it to send it as SMS or MMS. I don't think (maybe this has changed?) that that option is available immediately.

>I don't care how the message is sent, as long as it is sent.

Friend of mine has an iPad and an iPhone, often her iPhone data plan runs out so I send a message, it's blue and acts like it's delivered.... then I get a reply 6 hours later when she finally gets home and sees it on her iPad because the iPhone never received it because it didn't have a data connection.

Worst part is sometimes I know I need to send it SMS and it's a real pain trying to convinced iMessage to send SMS when it think the person is online with iMessage due to an iPad being left on somewhere. Can never remember which tap and hold dance actually gives me the SMS option.

This x1000. Plus when it silently decides to send emails instead, or you email as caller id address, or whatever it does. When my Dad asks "Why I'm I getting emails from text messsages" I don't know, don't care to know, and will respond by not using iProduct. Don't care if it's a setting or what, if you're going to merge different services then be explicit about it. Don't have time to figure out what you thought was best for me.

Is that Messages or is that the phone company that does that? Mine always come from an AT&T email account with some generic sender, like "aehcu2341@att.com".

Ya I've had those too. Think it's a separate issue but was never able to figure it out.

Tap and hold the message you sent over iMessage. Click send as sms.

Which is a totally undiscoverable feature.

Undiscoverable features is the new way to design apps! Look at Snapchat

'Swipe randomly in random directions from random screens until you see something you didn't know was there' is the new 'simple and direct'.

Not only is it undiscoverable, you don't even think you need to do it. Like others said- you take out your phone, send a quick message, put your phone away and do something, only to find out your message wasn't sent. Possibly because the recipient, unknown to you, didn't have data access.

I've been using Telegram for my cross-platform messaging and it works quite well. On OS X, it feels like a Mac app with embellishments, on Windows it feels modern and clean like a Windows 10 app should, unlike most apps on Windows 10 (including Microsoft's). On iOS it behaves like an iOS app, and who knows about Android but I would guess it's the same.

And yet, they also all behave the same. There's a consistency across all of the apps that makes it clear they're the exact same app, but there's also a consistency with the platform that makes them feel like native apps everywhere.

> Apple knew best with iMessage and decides it should invisibly merge sms and their own messaging.

Point taken, though I'll point out that iMessage makes it clear the distinction between iMessages (blue) and SMS (green), and provides an easy 'send as SMS' option.

Google's message integration, on the other hand, merged SMS and GTalk, with the result that I have several friends who would try to text someone to say 'hey where are you at', and it would send them a GTalk message to their laptop, which was turned on and signed in at home, rather than an SMS to their phone. Worse, I know one friend who couldn't figure out if there was any way to tell which one it sent, or how to force it to do one or the other.

In other words, Google did the same thing, but even worse.

> Result? A dozen times a day it would hang for 10 minutes or so trying to connect to data-unreachable imessage servers. Just send a damn SMS, which is what I wanted in the first place.

I haven't had this specific issue frequently, except when I had extremely unreliable data services (i.e. if I had a data connection but my network was broken, if I was on Wifi but my internet connection was down, or if I was in San Francisco), but I understand that there are definitely people who experience one of those issues very frequently.

> Everyone needs to just stop knowing best. 99.9 times out of 100 they don't.

And yet, your hatred is directed at the PoV that companies should respect an OS's conventions, yet who says that the OS knows best?

Well, if "best" is "follows conventions", then maybe the folks that set the conventions "know best"?

My point was the direct contradiction in statements there. Choosing the choices that a platform makes as the standard setter is arbitrary in a fashion, as why should those standards be the standard? There is no real reason why that makes it better.

Yes there is - it makes your app integrate better with the platform. Sure, the initial choices may have been completely arbitrary, but as an app developer you have to live by them now. Consistency is the reason. Users learn the platform and learn its standard UI paradigms. Should you choose not to follow them, your users will not be very pleased to say the least, often confused as to why your app doesn't operate like everything else on their device.

No, the whole point is that if everyone on the platform does it in one way that is "8/10 good" and there is another way that really objectively is "9/10 good" superior, then still the proper choice would be to follow the convention, as for the end user inconsistency is a much bigger drawback than the (comparatively small) improvement that can be done by doing it differently.

The operating system set a policy, and the user chose that policy when they chose that operating system.

Customers, who vote with their pocketbook.

True, the OS does not know best, especially in the case of Android. But the user benefits a great deal from consistency of behavior across all apps.

The bar for deviating from OS convention should be high, not only to preserve the user's sanity, but also because most users will tire of an app's design innovations, however brilliant, if those innovations impose a learning curve. Most users are not that dedicated to a piece of software.

It's not about "knowing best" in that case, it's about familiarity and reducing cognitive load.

When someone picks up a new piece of software, they have to figure out 1) does it solve their problem, 2) how do they use it, and 3) how do they solve their immediate problem with it?

When you have familiar UI elements, actions, terms, and explanations, you can speed along Step 2 for minimal effort. If you introduce new interactions and ideas, you're adding a Step 2b that will not pay off in the short term. It might in the long term, but not right now.

Every question, struggle, etc that you put in front of a user is one more reason to not use your software. Some users really NEED it and will overcome everything but that's not something to bet on.

Often it doesn't, but to me, that's not the point. The point is the consistency between disparate apps from different vendors and backgrounds is what matters for usability. That ability to know your way around nearly every app on first use, just from knowing the OS trumps most else. There's lots of things I use occasionally - I don't want to RTFM every time.

I dislike how Windows and Android do a lot of things, and how OSX does some. I'd still prefer devs respect that than "improve" on it. When I find a utility to add a dark theme or fix some niggle the OS does really badly - multi-screen perhaps, the standard stuff interoperates with it just fine. The clever "knowing best" program is the one that breaks it, or interoperates only partially. eg Photoshop has a lovely dark GUI now, but now doesn't get the multiscreen util's extra icons.

Let the user decide. They can add a multi-screen util or theme, or even simple things like xmouse and they should work everywhere, or they can just use the vanilla OS as they prefer.

EDIT: caseysoftware puts it rather more succinctly.

There are a lot of prominent counterexamples though:

- Blender (Blender's UI is pretty odd and entirely un-platform-like, but very efficient)

- Photoshop (case in point, Ctrl-Z; I'm very glad Adobe didn't follow conventions—Photoshop's Ctrl-Z behavior is good for Photoshop)

- Chrome (I guess it's standard on Android now, but when Chrome was released it was very unlike a typical Windows or OS X UI, and it was so good that everyone else adopted it)

- Vim

- Emacs

- Snapchat

- AutoCAD

Personally, I do prefer 40 unique UIs optimized for long term user happiness. That's kind of the Unix way: bc, awk, find, and dozens of other tools with different patterns that are geared very well to their niche.

Chrome - also interacts marginally with my multiscreen util. By breaking windows guidelines for topbar size the multiscreen icons get overwritten by tabs if I open enough. Workaround = have a somewhat larger top bar than standard or wanted. Still it no longer hijacks clicks to the icons which it did for quite a long while.

"so good that everyone else adopted it" Yeah, about that. FF copied Chrome wholesale, causing a couple of forks and quite a lot of anti-Australis outrage. Now there's a ton of hacky addons to restore FF to a relatively sane UI (none manage it completely). Will be interesting to see what happens when FF deprecates current addons and theming. That's actively user-hostile.

vi, emacs and AutoCAD are from long before GUIs, I don't think they count. vi and emacs made choices for 2400 baud terminals. Similar for the whole unix toolset - they were the standardisation being part of the OS. Much later we would argue sys V Solaris vs SunOS instead or CDE, vs KDE, vs OpenWindows, and start to consider GUI/OS standardisation.

Every time I want to do something on Windows I feel it, and don't do enough on Win to be bothered learning much powershell. I usually end up with a quick zsh script via network. Who on earth thought batch files were a good idea. :)

We're digressing though - GUI gestures in a slew of phone apps are rather different to a neatly focussed command line set.

Photoshop's overloading is sane, and I can't really disagree - I also prefer the dark theme. But it can, and does, break things.

All the examples you mentioned are productivity apps in which an optimal workflow which needs to be learned is more important than a familiar but not necessarily ideal workflow.

If I use an app 5 minutes a day, i want it to be familiar because I don't want to muck around learning how to use it. Whereas if I'm using an app all day to make money, I will assume the initial learning curve if it means increased efficiency.

Reply on news.ycombinator.com