Android: iPhone's Refugee Camp
While Android's open approach undoubtedly led some developers to pick it over the iPhone from the start, Apple's byzantine approval process and perhaps anti-competitive protection of its own apps—Podcaster and MailWrangler being two of the most prominent—have definitely driven some devs into Android's open arms, or at least made them stare longingly at it.

One such dev was the maker of the ridiculously popular Instictiv Shuffle app for jailbroken iPhones. Currently, iPhone apps aren't allowed to touch a user's music or iTunes functionality in any way. Instinctiv CEO Justin Smithline told us that "the minute we found out about the restrictions of the SDK...we started up an Android effort." Nevertheless it was clear in our interview that they loved the iPhone platform, using the word "amazing" more than once to talk about it.

Free But Not Equal
One of the original dustups around Android was that the 50 finalists in the Android Developer Challenge received early, privileged access to SDK updates that the rest of the developer community didn't get. While it makes sense that Google would want to fast-track Android's potential killer apps in time for the launch, it also goes against Android's atmosphere of openness.

It seems like there is some favoritism—whether it's toward specific devs or just toward the best apps is uncertain. TuneWiki is a finalist and one of Android's 10 most exciting apps. Amidst complaints about the lack of updates to Android's SDK until the recent 0.9 release and Google's secrecy, TuneWiki CEO Amnon Sarig told us that "I cannot say good enough things to say how [Google] treated us. They gave us whatever we wanted. They want us to succeed."

Since TuneWiki looks like it'll be a fantastic app, it's hard to argue with this—why shouldn't Google devote the most resources to the best and brightest, the stuff that'll make its platform shine? Logically, it should, given how much of the platform's success ultimately lies in the hands of developers. Depending on how you see Android's raison d'tre, that might be deeply troubling philosophically, on the other hand.

Nuts and Bolts
One thing that every developer we talked to pretty much agreed about is that coding for Android is not exactly warm robotic apple pie. While it's commonly assumed that Android development is done using run-of-the-mill Java, the developer of BreadCrumbz—a very cool image-based navigation app that's one of the 50 finalists, told us that the Android framework is actually "very different" from a regular Java stack, so that even "experienced Java developers still need to learn." TuneWiki's devs agreed that there's a learning curve, but both said that since it's still Java at the end of the day, it's a short one.

Instinctiv was more down on Android Java, compared to iPhone OS X, when it came to porting their app. When we talked to them before the release of the 0.9 SDK, they said that "Android is a mobile OS unlike the iPhone system, which is really kind of a desktop OS." Because of Java, they lamented that it'll be hard for Instinctiv Shuffle to do any really heavy lifting without bogging the system down, so they didn't think they'll be able to make it "as personalized" as the admittedly outlawed iPhone version.

Access to hardware appears to be much better than with the iPhone SDK, even though BreadCrumbz's Amos Yoffe says that Android "doesn't let you access the hardware directly, you go through Java APIs which are abstracted from the hardware." He still says that it's "pretty good." TuneWiki devs raved that "Android doesn't sandbox you like Apple does, so you have more flexibility." Apps run in the background just fine, battery drain issues aside. And conversely to this freedom, security policies and threats should be interesting (and maybe terrifying for nail-biter types) to watch develop, though at the start, Android seems to strike a good balance between security and freedom (insert current events political joke here).

Flexibility is a huge thing for Android. One of its strongest points—that it's going to run on a ton of phones with a rainbow of specs—might also prove to be one of its weak points, and perhaps the biggest challenge for developers. TuneWiki's Sarig said that since the Dev challenge only provided them with a single set of specs, no one's had to deal with the issue yet. It's definitely looming, however.

He admits that they're going to "have to scale back for less powerful handsets," though he doesn't know to what extent, since no one's seen the pile-of-rusted-bolts end of the Android hardware scale. BreadCrumbz's Yoffe says that "it's a bit early to say" if performance variance between handsets will be an issue, no one will really know "until we get our hands on real Android hardware." The G1's hodgepodge of interface methods—touchscreen, QWERTY and trackball—perhaps not so coincidentally gives developers a chance to experiment with multiple ways to interact with their app using a single device, though having to account for them all necessarily adds layers of complexity and consideration to creating apps.

Android vs. iPhone: The Final Battle
Pandora CTO Tom Conrad, who famously said "I need Android like I need a hole in the head," actually takes a more measured approach to the platform war. He told us that "Generally, when I look at Android and the challenges we faced bringing Pandora to handsets," it doesn't seem to solve them. "It just adds another one to the mix."

Critical for Android's success is an easy-to-use app store and the killer apps to stock it. Conrad noted that while Pandora had been on a number of low-end phones for over two years, within 24 hours, their iPhone app had surpassed all of those users combined. They are currently taking a wait-and-see approach with Android, though he stressed that "absolutely, we want Pandora to be everywhere there are listeners." TuneWiki similarly wants to achieve multi-platform ubiquity, though they're much more juiced about both the iPhone and Android, saying, "We love them both."

Whoever wins, it looks like the carriers will lose. Every dev agreed that the iPhone sparked a revolution that is changing the way US carriers operate. Android is a part of that now, and the two, even locked in competition, will push that revolution further. In that sense, at least, we all win something.

Hokay, the iPhone 3Gness makes browsing a whole 2.4x faster than EDGE in Apple's test. (One thing that we can't explain: Why did Apple chose lonelyplanet.com for its performance benchmarks?) The 3G goodness is real: We've been conducting our own testing of AT&T's HSDPA in the New York area (including suburbs) and it really is faster and more readily available than Verizon's EV-DO 3G network. (For a quick primer on different kinds of 3G like HSDPA vs. EV-DO and other mobile terms, click here.) But the iPhone 3G is rated for 1.4Mbps, a nice clip but not the 3.6Mbps downstream that AT&T's HSDPA is capable of. (The carrier loves to brag that it'll have 7.2Mbps by the end of the year.) So why not crank up the iPhone to those better data rates? Turns out, according to AT&T people we talked to, 1.4Mbps is the capped bandwidth for all mobile smartphones on the network for a few reasons.

(UPDATE: AT&T is saying they're not capping the phone at 1.4mbps, but that's what its capable of doing now, due to factors below. There's no difference except intent, and AT&T is careful around words like "Cap" these days.)

A major one is battery life—the faster you burn, the faster your battery dies, so going full steam at 3.6Mbps would cut you well short of that nice round five hours. A second one is cell site congestion and backhaul (carrier-speak for size of the wired dataline that connects cell sites to the actual telecom infrastructure). While everyone at AT&T, from the top down, is adamant that AT&T is "comfortable" with their ability to meet the huge data draw once 3G iPhones hit the streets, it's not like the pipe is unlimited.

AT&T wasn't able to give a breakdown as to how many of their towers have fiberoptic pipes as opposed to slower copper T-1 lines. Nor could they say how quickly they could add capacity to a site that is pummeling their demand expectations, since it varies from site to site. Ones in dense urban areas are loaded up with more backhaul and can handle more users than one closer to the edge of their 3G footprint. Still, generally speaking, more users on a site means more congestion, so if you're slurping from a site that's really slammed, it will be slower. As with all radio technologies, proximity also matters. (Hint: For the absolute fastest speeds, wait until 3am and then go sit right next to your favorite cell site.)

Something we missed, or you still wanna know? Send any questions about 3G, GPS, G-spots or anything else to tips@gizmodo.com, with "Giz Explains" in the subject line.