I spent some time looking through the just-emerging web OS technical documentation, and it quickly became apparent that Palm’s approach is radically different from both Android’s and Apple’s. Since they’re all here at more or less the same time, running the same Web browser on roughly equivalent hardware, this represents an unprecedented experiment in competitive software-engineering approaches.
“Equivalent Hardware”, You Say? · Yep. They all have 3G, WiFi, GPS, Bluetooth, bright high-res screens, pleasing form factors, you name it. Apple has historically had more design mojo, but on the other hand they suffer from a certain institutional tight-assedness that has led, for example, to the iPhone having only one button while my G1 has a lovely trackball, a super-handy “Back” button, and a slide-out keyboard; this is the same culture that resisted multi-button mice for so many years.
Apple may hold a bit of an edge if they can defend the patent fence around multi-touch, but that’ll be a hard perimeter to maintain.
Given the proliferation of lean, mean, and smart handset manufacturers, I just don’t see any one party being able to grab and hold a hardware lead.
Software From Different Planets · It’s like this:
| Language | Framework | Notes | |
|---|---|---|---|
| Apple | Objective-C | Cocoa | Old-school object-oriented language compiled to the metal; general-purpose UI framework with roots reaching back to NeXT. |
| Android | Java | Android | Java language, custom VM, built-from-scratch UI framework aimed at small-form-factor devices, fairly abstraction-free, based on “Actions” and “Intents”. |
| web OS | JavaScript | “Mojo” | All Web technology all the time. Innovative and visually-impressive “card”-based UI. |
There are some things missing from the table above: Blackberry, JavaFX Mobile, and Windows Mobile. Blackberry, because they don’t make much noise about developer tool and what one does hear isn’t that encouraging.
JavaFX Mobile has just now released and I haven’t had any hands-on yet, but I’ll totally give it a try and report back. As for Windows Mobile, my perception is that they’re just nowhere near a leadership role in terms of user experience or developer pull.
The Experiment · An unstated premise of this experiment is that developers are important. If the hardware is mostly a wash and everyone’s got the same browser, the big differentiating factor here is going to be the quality of the apps.
Now, I can make a good argument, from theory, as to why each of the approaches featured in the table above is the right way to go for the mobile space. But I don’t have to, because we’re all going to get a ringside seat while the world runs the experiment, then we’ll know. Apple’s 2-year head start is material but not, I think, decisive.
This is going to be fun.
Contributions
Comment feed for ongoing:
From: Avi Bryant (Feb 18 2009, at 12:26)
Developers are important, yes, and I suspect the quality of the apps is going to have more with the quality of the developer community than the technical characteristics of the platform. The shareware/indy developer community on the Mac has always been exceptional, and I predict that iPhone apps will always be significantly more polished than the alternatives not because Cocoa is better, but because Cocoa developers are, as a group, more obsessive about UI quality than their Java or even Javascript counterparts.
[link]
From: Drake Anubis (Feb 18 2009, at 12:32)
Hardware is misspelled in the large sentence there.
Also its my understanding that the palm pre can preform certain tasks dramatically faster than the iPhone or G1 (for one, it's processor is a superscalar). I'm also under the impression that the graphical processor is much more capable.
[link]
From: michael thorne (Feb 18 2009, at 13:00)
HTML & javscript/ajax VIA "Richard Feynman's 'turtles all the way down'".
The web's the way it is because HTML & javascript are "simple" and easy to hack
Give that to the "kids in the garage" so they can work their mobile devices the way they want and not rely on the "Corporation" and they will make it happen.
I think the "webOS" is the way to go.
Sorry for all the quote/unquotes.
mxt
THINK
think different
Think Open Source
[link]
From: Duncan Davidson (Feb 18 2009, at 13:40)
I think you're absolutely right. Instead of having big heads pontificate about the "correctness" of a particular approach, we're gonna see this one work it's way out in a real world way. Which is interesting. After all, the traits that seem to matter most might not matter at all in practice. The market will give us a bit of insight.
I'm personally looking forward to at least playing with a Pre. Boo Sprint here in the US and the like, but...
[link]
From: stand (Feb 18 2009, at 14:39)
I just took a look at the Blackberry development situation. Ugh! Looks like non-Windows developers need not apply. Someone please tell me I'm wrong on that particular point.
[link]
From: Bob Aman (Feb 18 2009, at 14:48)
My money's on the iPhone winning, but I do intend to hedge my bets eventually. I'm writing an iPhone app, but I plan to write a port for Android when I've got the time, and assuming the iPhone app sells well enough to justify it. I haven't had any hands-on time with either of the other two devices yet. I'm most curious about the Pre. How do they handle persistence and database support if everything's JavaScript and HTML?
[link]
From: Andrew Shebanow (Feb 18 2009, at 15:46)
Bob - from the WebOS book chapter linked by Tim above:
"Mojo supports the HTML5 database functions directly and provides high-level functions to support simple Create, Read, Update or Delete (CRUD) operations on local databases. Through these Mojo Depot functions, you can create a local database and add, delete or retrieve records individually or as a set. It's expected that you'd use databases for storage of application preferences, or cache data for faster access on application launch or for use when the device is disconnected."
[link]
From: JulesLt (Feb 18 2009, at 16:08)
I can't see the iPhone winning in the long term - and I say that as someone who has an entirely Apple based setup at home.
The simple reason is that I can't imagine a world where 70% of people choose an iPhone for their mobile (i.e. an iPod level of domination).
On the other hand . . . Apple were criticised when they bought time by saying HTML+JavaScript was a good enough developer platform. Perhaps because it was clear that Apple's own apps were not???
It's an interesting decision on Palm's part anyway - it definitely means no Super Monkeyball or other performance intensive apps - but aside from games, I suspect most successful iPhone apps could have been implemented as web apps.
The ideal situation for consumers is for at least 3 dominant players to stay in the field - so long as we don't let our data get locked to the platform, software can come and go, particularly if it's cheap. People don't get locked into console platforms, despite their software libraries.
And ideally I'd like to see all of the next generation implementing a wider choice of programming languages - i.e. PyObjC/MacRuby on the iPhone, and Python/Ruby as alternative browser languages to JavaScript in the WebKit browsers, similar to IronMonkey and the DLR.
I know it's just snobbery to rag on JavaScript and that with careful use of features you can write good clean code . . but as someone who has been used to being able to choose languages since the mid-80s, being forced to use the vendors built-in language feels like a step backwards to the earliest 8-bit desktops.
One other plus for the iPhone - strong desktop client integration.
[link]
From: JulesLt (Feb 18 2009, at 16:27)
One small point - the iPhone's UI framework is possibly as built-from-scratch as Android - i.e. the iPhone uses UIKit rather than the same AppKit/NextStep classes as MacOS.
No idea how much the actual implementation varies (i.e. many of the classes have the same names bar the prefix).
3 more candidates to throw in :
Nokia - they're up to something with open sourcing Symbian and purchasing Trolltech. I think they have the best chance to create a platform that actually rivals the iPhone - i.e. C++ for close-to-the metal speed, Qt as modern UI toolkit - plus with NGage there is some experience with creating a gaming platform, although I'd call that a failure.
Then there's the cross-platform choices :
JavaFX - big unknown for me - there are a lot of Java developers out there, but are there a lot of great client side developers?
Flash platform - if they can get it to perform reasonably on small devices.
[link]
From: Drew (Feb 18 2009, at 17:27)
Having worked with the iPhone dev environment I wouldn't call it general purpose UI. It's very much mobile focused from soup to disclosure arrows. That being said Android has a great answer to the background app problem so I hope I'm not coming off as too partisan in all this.
[link]
From: Parand (Feb 19 2009, at 01:11)
Developers are important, but number of users is much more important. I personally know several developers building iPhone apps not because they like the development environment (most of them hate it), and not because Apple provides an easy path to publishing apps (they make it ridiculously difficult in some cases), but because the iPhone has such a large market share.
I'd much rather be using Java than going anywhere near Objective-C or Cocoa, but until Android market share picks up it ain't going to be happening.
On the other hand, Gina Trapani dumped her iPhone for the Android partially because of the development environment (http://smarterware.org/184/why-i-switched-to-android-from-the-iphone), so maybe there is hope.
[link]
From: Parveen Kaler (Feb 19 2009, at 09:23)
I also would not call Cocoa Touch a general purpose UI framework.
Core Foundation, Core Animation, Core Data, etc are very much like the Cocoa framework on the Mac.
But, UIKit on the iPhone is very different from AppKit on the Mac.
The model-view-controller design pattern is very consistent across both frameworks, though.
[link]
From: Derek K. Miller (Feb 19 2009, at 10:03)
There's another factor here, where only Apple has a current example: as a developer, once you make it past the approval hurdle (no small barrier, but still), you have an easy, convenient, consistent way for people to give you money for your product on a whim. All that infrastructure is taken care of in Apple's 30% cut.
In principle, many free software/open platform advocates might hate that. But there are developers like like Pangea who have made many times more money on iPhone apps than they ever did on Mac or PC software. So some are moving to all-iPhone development, because that's where the money is:
http://themacgamer.com/2008/09/09/brian-greenstone-i-have-no-plans-to-do-any-more-mac-games/
Android, WebOS, Symbian, and others could offer that too, but right now there's no sign of a similarly seamless purchase experience, either for users or for developers. That's the iPhone's current secret weapon.
It may be that a ton of cool open-source stuff appears for Android and WebOS, but it will likely be the kind of software open-source programmers need. The iPhone App Store provides an incentive for developers to build the kinds of apps that users pay for because they like them. From what I see, those are two different categories of software.
[link]
From: Marius (Feb 19 2009, at 14:38)
Nokia should not be underestimated. There are over 100 million smartphones out there that are based on Symbian.
[link]
From: Derek K. Miller (Feb 19 2009, at 17:23)
Another interesting perspective from Joe Wilcox at eWeek, via John Gruber at daringfireball.net:
http://blogs.eweek.com/applewatch/content/app_store/theres_an_app_for_that.html
"Apple’s platform will grow stronger and maintain huge advantages over competitors as long as there continues to be one iPhone OS version for all handsets from all carriers. [...] Rather than there being multiple mobile OS versions, further fragmented by carrier distribution, Apple controls and distributes the updates."
One way to kill developer and user interest is a forked OS that may or may not run your applications reliably, or at all.
[link]
From: Tony Fisk (Feb 19 2009, at 17:28)
"As for Windows Mobile, my perception is that they’re just nowhere near a leadership role in terms of user experience or developer pull."
Shortly after reading this, I saw a news article about a top Telstra executive having his mobile phone pinched. Apparently, it contained a beta version of the new Windows Mobile software (all highly confidential)
http://www.theage.com.au/news/technology/biztech/pickpocket-swipes-secret-telstra-phone/2009/02/19/1234632937082.html
An interesting snippet, which doesn't detract from your thesis.
I suspect the 'winner' is whoever can leverage the bottom end of the market (as in solar powered mobiles.. you may not need it, but a lot of people do! Far more than there are careless Telstra executives!)
http://www.theage.com.au/news/digital-life/mobiles--handhelds/articles/solarpowered-mobile-phone-could-be-a-big-hit/2009/02/19/1234632945177.html
This is where it could get interesting. If it were a matter of software only, then I would say an open source offering, such as Android or Symbian, would win out. However, it will require a smart and dedicated operation with funding to get a low power (ie solar) smartphone that is accessible to the third world
[link]
From: JB (Feb 25 2009, at 19:59)
You make an interesting point, but I'm not going to get my hopes up for the quality of this experiment (much as I'd like to see it play out). It seems that what happens with these things in general is one of the competitors makes a Big Mistake: rewriting the whole thing from scratch, say, or shipping some horrendous hardware or making a really unlucky arch-move. The probability that 3/3 platform developers actually execute a strategy this complicated without an exogenous "Big Mistake" seems near zero to me. </handwaving>
[link]