Typical poor TechCrunch reporting...<p>"We’ve spoken with a number of high profile Android application developers. All of them, without exception, have told me they are extremely frustrated with Android right now. For the iPhone, they build once and maintain the code base. On Android, they built once for v.1.5, but are getting far less installs than the iPhone."<p>Is he suggesting getting far less installs is related to maintaining a single code base? Really?<p>"They say they’re going to have to build and maintain separate code for various Android devices. Some devices seem to have left out key libraries that are forcing significant recoding efforts, for example."<p>I have experienced broken APIs in Android OS updates, but I've never experienced a missing library between devices running the same Android OS version. If he's going to make such a claim, he should back it up.
A few years back, I worked on some software for phones and PDAs running Palm OS. To do something as fundamental as send and receive SMSes, I had to use a completely different API on each supported device.<p>It was a complete nightmare and really depressing work, writing the same thing over and over again. Google really need to avoid a similar situation on Android.
J2ME didn't fracture because it was open source - that didn't happen until 2006 - it fractured from the same issues bensummers points out below with Palm: too many differences in implementation between different manufacturers and devices (including bugs).<p>Phone B supports 1-bit alpha, phone B supports 8-bit alpha, phone C has a bug where alpha is corrupted in images under a certain size. Once your code starts looking like a nest of if(NOKIA||(MOTOROLA&&!MODEL910)) you're doomed.
It's definitely something they're aware of. I know because I personally asked some of their developer advocates what their strategy was at their developer meeting in Munich, and followed up elsewhere.<p>I didn't get an answer, but they're not stupid, it has to be something they're thinking about. Now, let's just hope they don't screw it up.
Does Arrington <i>ever</i> cite sources or quote people or name names? "I talked to some people that said some stuff." Wow, how valuable. I have interesting conversations with The Voices in my head too, but I don't write about it...<p>Some thoughts: Desktop software suffers from this problem, everyone's configuration is slightly different. You just have to test the cases, or simply say, "this software doesn't work on your phone". (If the software is Free, then this isn't a problem, as an interested user can just fix it himself. This isn't even a possibility on Apple's stack.) Video game developers somehow manage to make their software work on PS3s, Xbox 360s, OS X, and Windows... and phone developers are complaining that they can't support both Android 1.5 and Android 1.6? Wow, OK. This might indicate code that is too low-level; Android's own API is not enough of an abstraction layer for the average use case. (Wouldn't it be nice to write apps once and have them work on every phone platform?)<p>I have not done any significant Android development, but so far, everything has worked as expected. The simulator works like my phone. As a user, all of the apps I've installed (via the Market or otherwise) have worked just fine.<p>So I think problems here might be overstated, except for one -- there is no money in writing mass-market mobile phone apps. (Thank Apple for that one.)
The mobile phone operating system battle is going to be fascinating over the next few years. I'm sure it will be studied in business schools for the the next 40 years. The stakes are extremely high and the market is global. Hardware, software, and carriers all play an important role. Overall, consumers will definitely win.
<p><pre><code> - [Android will be] second most popular mobile OS after Symbian
- [Android] operating system is free (unlike Windows Mobile)
</code></pre>
Hm, interesting logic. Seems like Microsoft-bashing infiltrated very deep into thought process of many authors.