The “iPhone and Android” session was presented by Julios Barros, who quickly became more of a moderator as the session changed into a lively and informative audience discussion. This post focuses on the session conversations rather than the slides. I’m going to refer to one particularly knowledgable/world-weary attendee as “Mr. Dude” because I forgot to ask his name when I thanked him afterwards for his comments. This was the only technical session where I didn’t have enough background (on Android) to follow in detail, and it was an iPhone-centric conference anyways, so take the notes with a grain of salt and let me know if anything needs correcting.
Kenji Hollis' neat trick.
Alternate marketplaces for mobile phone software…
General lack of enthusiasm in the room for alternate mobile software marketplaces: Handango, Slideme.com, AndAppStore.com, Mobihand.com, etc. Reasons cited: lack of momentum, historically low commercial success rate, and costly paperwork/certification barriers to entry. While various carriers with various phones may support Android, Mr. Dude noted that carrier certification may restrict access to certain features in third party software. In other words, your users could download your software from an alternative marketplace, but without certification, successful installation doesn’t guarrantee software would run on specific carriers. On the other hand, Apple won’t even let your app onto the store if they find you using unauthorized capabilities of the device.
Open source license… what does your mom think?
Android’s open source license appeals to developers, but mainstream consumers don’t think about it. Keith Peters got a laugh from the room noting that “your mom doesn’t care about open source.” Interesting hypothesis came up that Apple’s success with a highly proprietary approach may have pushed all the competitors to embrace the opposite extreme of open sourcing as much as possible in the hopes of catching up.
Mr. Dude insisted but declined to elaborate on why people should read the Android SDK license more carefully to determine its degree of open source-ness. I am not a lawyer and not familiar with Android at all so I’ll refrain from commenting here. I did find citations 23 through 28 on Wikipedia’s Android page with more to say on this subject.
Portability and technical notes…
Two groups debated theoretical vs actual costs of porting Android. Some suggested that open source meant we might see Android on a much wider variety of vertical niches soon, while Mr. Dude and others cautioned that porting required significant investment from manufacturers. The Open Handset Alliance defines a baseline of capabilities followed by Android, and CCPP support means you can get a list capabilities of conforming phones. However, the desktop PC doesn’t support CCPP, so porting mobile software to heterogenous desktop hardware might require another way to detect capabilities. Also, Android’s UI framework ultimately talks to a native layer called the SKIA Graphics Layer, the same graphics engine used for the Google Chrome browser. Porting may require heavily modifying or replacing the SKIA layer.
Miscellaneous technical notes: Mr. Dude noted Google’s own Android JVM implementation is bound to ARM processor instructions, though I’ve since found news of a JVM for x86. A few folks concurred that memory copying is 100x slower than iPhone… I found this perf test with some corroborating numbers. Google has not published any JIT compiler, it’s all JVM interpreted right now. Someone with experience in both SDKs noted that if you’re a java desktop/j2ee dev, you’d be comfortable with Android. If you come from J2ME, you might feel more comfortable with the concepts and architecture of the iPhone SDK.
The G1 phone has a built-in compass, and Kenji showed me a few demos afterwards where having a compass lets you do much more useful things with the accelerometer and GPS. I’m keeping a closer eye on the Android after this session.
Business considerations…
No client buzz about android detected by attendees so far. Of course, this is an iPhone conference so the attendees’ client base probably skews towards iPhone projects. Some mentioned the strategy of building iPhone apps for clients first, and only for Android if requested. In such a case, they’d outsource the android porting work with the functioning iPhone app as the reference implementation. One caveat is that an iPhone-to-Android port may require significant redesign to provide a gracefully degraded or improved experience, depending on screen dimensions and device capabilities.
Some attendees voiced concern that they didn’t perceive as strong focus, vision, responsiveness from Google as Apple, with less public awareness too. On the other hand, there was lots of excitement about Android’s suitability for customized, vertically branded hardware scenarios.
No success stories for Android yet. Someone asked if anyone had written an Android application identical to an iPhone application? Silence. Mr. Dude and others said iPhone apps would need major rewrite, big pain in the neck. On bright side, Google’s looking for Android showcases to validate the platform, so there’s opportunity to get a big marketing push.
Unlocked G1 developer phone…
One of the speakers, Kenji Hollis, mentioned a quick, legal way to get an unlocked, internationally compatible G1 developer phone for around $400 without paying a monthly fee. I caught up with him afterwards to get details. After we borrowed a paperclip from the front desk, I took notes + photos of how he swapped an AT&T 3G SIM into an unlocked G1. He’s cool with my posting the step-by-step details, which I’ll do later if I have time.
Grain of salt…
Again, these are notes from an Android discussion at an *iPhone* conference, so take it with a grain of salt.
Yay! I’m glad I was able to make an impression with the G1. I was happy that Julio gave a quick presentation on the G1 vs. iPhone, and I feel that there were good questions asked at the roundtable of sorts. I didn’t actually think anyone was interested in the G1 unlock, but I’m glad that I was able to demo it.
Thanks again!