Tweetie used to be just a very popular iPhone client for Twitter. Now it has moved it to the desktop and it’s going to be one of the most popular desktop clients out there. Why? Because they focussed on solving the problem without getting caught-up developing a desktop application, and as it turns out, their “mobile” solution worked very well. So why change it?
For the last five years we’ve seen applications migrate from desktop to mobile. We’ve seen developers struggle to replicate the “desktop” experience, the desktop UI, the fluidity, and for the most part it has been painful stuff.
What’s interesting about Tweetie moving in the other direction, is the trend that it represents. Tweetie is the best application for Twitter, period. Good iPhone applications aren’t just on a par with their desktop equivalents, they’re better. This is the benefit of constraints.
There is no “desktop”
The old metaphor of a businessman’s desk complete with a leather chair, a rotary dial phone, files, folders, filing cabinets, clipboards, and dictaphones is dead. We work with computers now. We don’t need this metaphor anymore. It doesn’t make sense.
For every few Tweeties there’ll be the occasional TripLocator, but overall I believe that iPhone is showing designers a way forward. A way without “File->Save” or endless context menus. A way where everything you need is placed at your fingertips and there is no “read only” mode.
Consider the follow snippet from the iPhone UI guidelines…
Save data when possible in case the application exits? Now when wasn’t that a good idea?
It may take something like the iPhone to make us realise that we can finally throw out old conventions & metaphors and start looking for the new ones. Ones that make sense.
19 Comments
The same is true for web-apps. We can learn a lot from iPhone design. Design with constraints and you have to think that much harder.
Posted by Paul M. Watson at 10:12 am on 21 April, 2009.
I think
The ’save’ issue is a complex one. A lot of the time ’save’ is not desirable without some sort of versioning control. What’s interesting is that 30 million people ‘coped’ on iPhone/iPod touch without explicit Save, without Copy and without Paste. But on a device where the paradigm is that you could be disturbed at any time, it makes a lot of sense. And people have coped, even thrived. We’ve taken on a UI convention due to working under a constraint and lived to tell the tale. That’s a lesson right there. Everyone built save into their apps (in Cocoa, it’s free), because we thought we had to.
Tweetie is doing well because it’s better than the alternative which was either Twitterrific, the Twitter web interface or installing an AIR app - it’s not perfect by any means - but the alternatives were truly awful.
Is the world now ready for a difference in interface design? We’ve used the current one for 25 years+ now. Everyone is talking about 3D interfaces, task based interfaces, wizards, whatever. I think Touch is the same fad. We’re at least considering interfaces that don’t have a Start button which is a success in some form - change in the last 15 years!
There’s definitely a drive towards task-apps - discrete applications which service one purpose (as opposed to task-based which hand-hold you through trivial tasks).
Posted by mj at 10:36 am on 21 April, 2009.
@MJ - I agree the alternatives were truly awful, be it on desktop or iPhone, or web based.
In terms of “is the world ready” , it’s tough to tell, and I’m not sure it matters. People are always hesitant to change. There is nothing more usable than what you can already use etc.
I believe that Touch is a type of fad. I hate the whole “Tap is the new click” argument, it’s totally ignoring the context of click.
But I do believe that we’ve a lot of baggage in UI design that we can shed. And the iPhone has helped us shed a lot of it, and raised the bar while doing so.
In terms of discrete apps that serve one purpose, it’s funny. UNIX heads have been making that argument since before I was born, is this just another case of Apple dragging them to popularity, kicking and screaming?
Posted by Des Traynor at 10:49 am on 21 April, 2009.
I wouldn’t confuse an explicit save with an implicit one. If you hit the home button while typing an email on your iPhone it implicitly saves it. On returning to Mail.app it restores the state. Nothing was sent or saved, just an app. state remembered. You don’t need version control for that or any implicit save.
And I wouldn’t call touch UI a fad. If we had the technical capability back in 1963 we might have skipped the mouse all together. Touch is a more natural interface than the disconnected mouse idea (move this bit of plastic and the pointer on the screen three feet away moves.) However we have lived with UI input disconnect (keyboard, mouse, stylus) for a few decades now so any alternative will meet resistance.
The iPhone UI is more than just touch though, it is a focus on task based UI.
Posted by Paul M. Watson at 10:57 am on 21 April, 2009.
> In terms of discrete apps that serve one purpose, it’s funny.
> UNIX heads have been making that argument since before I was born,
> is this just another case of Apple dragging them to popularity, kicking
> and screaming?
The Palm webOS seems to embrace this much better than the iPhone OS. iPhone apps are still very disconnected. If you want the compose mail view in your app. you have to recreate it completely or send focus over to Mail.app which does not then return focus to your app. Apple have only surface a few discreet components like maps and video views.
I’d love to see deep-linking in the iPhone OS with real return links.
Posted by Paul M. Watson at 11:01 am on 21 April, 2009.
Touch UI isn’t a fad, I was being extreme. But whilst it’s definitely intuitive, it’s not necessarily the right way to do things. There is a lot to be said for minimal input, maximum output. You don’t want to spent your working day on a touchscreen application with a big monitor, it gets very tiresome very quickly.
Posted by Des Traynor at 11:06 am on 21 April, 2009.
@des - Remember that NeXT was a big gang of OO-fanatics who were also complete UNIX-heads. In 1997, when Apple was taken over by NeXT, it changed a lot of their priorities (for the better as it turned out).
@paul - I think Touch is a fad and it will go alongside desktop, wizard/assistant, depth and other metaphors to help us decide the interface of the future.
The difference between in ‘implicit’ versus ‘explicit’ on the iPhone is that when I press the Home button, I see that function in my head. I am executing something and the OS knows what to do, it’s explicit to me.
When a call comes in, or a UIAlert pops up, there’s not the same amount of user control. And that’s the only UI bugbear. There are a heap of games which react poorly to a scheduled alarm or an incoming text. This ’save state’ may be explicit to the phone but it’s very much ‘implicit’ to me.
Posted by mj at 11:08 am on 21 April, 2009.
@Paul - the lack of Return is something that bugs me and they’ve gone a little way towards solving the Mail issue with iPhone OS 3. There’s a mail ’sheet’ which you link into your app and Mail is permitted to be backgrounded so your app never really loses focus.
But yeah, it took to version 3…
Posted by mj at 11:11 am on 21 April, 2009.
That is likely because touch has just been added to existing desktop UI and often on vertical screens. You have to reach up and precisely tap the File menu, strenuous and exacting for sure. Flicking a mouse is always going to be less effort than even a horizontal touch UI.
But maybe a bit of extra physical effort to work with more tactile interfaces is a good thing. Put the wireframe flat on the desk and let 2 or 3 people reach out and turn, prod, grab and pull it. Let the graphic artist stroke on colour and form. Let the car designer work with his hands on his digital creation. These actions are deeply ingrained in our evolution. Standing up and flexing your muscles around an idea in front of you might have the same benefits as standing up with your co-workers at a whiteboard. Getting away from sitting for hours hunched over a mouse and keyboard.
Hackneyed but the projected interfaces shown in Iron Man were interesting. It forced the character to interact with what he was creating in the same way users would. He had to reach into his design, highlighting usage deficiencies of the design.
Posted by Paul M. Watson at 11:15 am on 21 April, 2009.
> What’s interesting about Tweetie moving in the other direction, is the trend that it represents.
If you’re claiming this is a trend you need to give more than one example. A single data-point is not a trend.
> Good iPhone applications aren’t just on a par with their desktop equivalents, they’re better.
This is a pretty sweeping statement and I don’t believe it’s true in general. Other than tweetie, which iPhone applications are better than their desktop equivalents?
Posted by Aidan Finn at 11:16 am on 21 April, 2009.
> which iPhone applications are better than their desktop equivalents?
Clock.
Calculator.
Maps. (vs. maps.google.com or Google Earth even.)
iPod (vs. desktop iTunes.)
YouTube (vs. web YouTube.)
Skype.
WhatTheFont.
App Store. (man, the App Store is awful in desktop iTunes.)
And I really like the way multiple pages are handled in mobile Safari over desktop Safari.
Posted by Paul M. Watson at 11:25 am on 21 April, 2009.
@Aidan
1) The trend I refer to is that desktop and web apps are copying iPhone style interactions, or “mobile” interactions if you prefer. In that they’re becoming single purpose context aware apps as opposed to everything apps. In addition, people are copying certain controls and widgets from iPhone apps, along with their primary navigation style. ( stuff like http://www.aboone.com/?p=34 and http://widowmaker.kiev.ua/checkbox/ is becoming common in apps)
As time goes by, I think we’ll see the influence of the iPhone on more and more applications, desktop, web and beyond. http://www.slideshare.net/stephenpa/inspiration-from-the-edge-new-patterns-for-interface-design?nocache=3170
2) The Harvest iPhone client is far better than it’s desktop counterpart, as it focusses on one application only, namely entering time. The Facebook iPhone App is more effective and quicker to use than the Facebook website. These are my subjective opinions, time will tell.
Thanks for your comment.
Posted by Des Traynor at 11:26 am on 21 April, 2009.
The implicit vs explicit save is an interesting one. One of my colleagues had this conversation with a Microsoft programmer on the OneNote team. They have implemented an implicit save on that desktop app, and it doesn’t sit well with some users (notably, my colleague) as the save convention is so ingrained in the desktop environment. I was on the side of Microsoft on this call,though, getting rid of Save in OneNote was a good move in design space.
Looking into the history of explicit Save, you can see that the convention comes solely from a hardware constraint: the fact that a computer’s memory is split between a persistent medium (a hard drive) and a volatile medium (RAM). What the Save command really means is “store this on the hard drive now”. If there was no distinction between volatile and persistent memory in machines, Save would have been unnecessary. Of course, this distinction could and should have been abstracted away in the software through frequent saves, this is simply an example of the implementation mental model of interface design.
Personally, I hope that Save is killed of through better programming as the anachronism that it is, and that this change is accelerated by platforms that don’t have the RAM/Hard drive distinction, like the iPhone or laptops with Flash drives. The new convention will catch on soon enough.
Posted by John at 12:52 pm on 21 April, 2009.
Was just discussing this with my boss - who, like me, is an old grey hair who remembers a Save function also being known as a “make the tea” function.
While we’re dealing with small interfaces and small files, Save can be done away with. When we’re dealing with workstation interfaces and workstation work (like editing a 2 GB image), we may have to take into account the need to Explicitly Save. Though with more and more cores being added, perhaps an Implicit Incremental Save is the future.
SSDs are, at the moment, no different to hard drives in terms of access speeds but like Mac OS X wasn’t built for the shipping hardware of the day (it’s slow…it’s laggy), the UIs we’re looking at should be built for the hardware of tomorrow.
Posted by mj at 1:13 pm on 21 April, 2009.
That’s a fair point, mj, not every situation is amenable to incremental and continuous save. But most are, and I agree with you that we should be looking forward to tomorrow’s hardware.
Posted by John at 5:47 pm on 21 April, 2009.
Yes, good points. Certainly I don’t want Photoshop auto-saving changes to a design file I am toying around with. But I do want it caching changes to disk in case of a crash so that on restore I can choose to get those changes and get the last version of the file. If it auto-saved to the file, creating versions as it went, you’d end up with a chaotic and useless history. You wouldn’t want subversion auto-committing either.
Good discussion guys.
Posted by Paul M. Watson at 5:27 am on 22 April, 2009.
Hi Des,
great post (as usual).
If it is a trend, and I hope it, another reason it works is that in designing for mobility (not mobile technology) it focuses the design on what’s most important.
This approach when applied back to the old fashioned desktop or web app works because perhaps designers and their clients have learned the importance of focus.
Take mobile banking, apart from below what else needs to go in my mobile design:
How much do I have in my account?
I want to transfer money/pay a bill quickly
I want to see the last 5 outs of my account
Now, why don’t I see these upfront on a dashboard when I login via the browser on my MacBook Pro?
Posted by Lar Veale at 7:26 am on 25 April, 2009.
Hey Lar,
Thanks for your thoughts.
It seems we agree. I wrote about Online Banking and the Benefits of Contraints before: http://www.contrast.ie/blog/the-benefits-of-constraints/
Cheers,
Des
Posted by Des Traynor at 4:54 pm on 26 April, 2009.
It would seem that iPhone App designers are forced by constraints and custom to implement the UNIX philosophy: simple tools that do one thing and do it well, while tolerating interruptions. That such applications can offer their “one good thing” for free or for a few bucks is (perhaps) training users to think of a software ecosystem in a new way.
Think about the difference between most iPhone apps and what most end-users think of as “software”: something like a commercial Office suite. For most people, this is the only software they buy. It’s very expensive, and there are tons of features that most people don’t know about and can’t use (but can’t be eliminated), it’s complex to navigate and there are huge implementation issues around such simple things as saving state. If software users get used to the idea that an application can do one thing well, perhaps they will better appreciate where UI/UX/IA design falls into the utility curve.
The web isn’t the only place we’re seeing this trend, of course, but it seems to be a good indicator for the future. Perhaps if we could think more about this kind of building-block simplicity, our software would be less frustrating and less fragile.
Posted by Breandán at 9:47 pm on 15 June, 2009.
1 Trackback
[...] wrote a lovely post about the death of the ‘desktop metaphor’ over on the Contrast [...]
Posted by Tasty morsels for lunchtime enjoyment at 10:42 am on 22 April, 2009