More Flash Thoughts

There's been a lot of interesting discussion around the whole Flash CS 5 generating iPhone Apps thing. None of it has changed my opinion that it's in Apple's best interests and ultimately the best interests of iPhone users for them to take steps to stop Adobe from moving forward with this, or at least for Apple to set some boundaries with Adobe and phase the program in over time to avoid overwhelming the App Store infrastructure. The problem is multi-faceted, which can be confusing. To me, it seems to break out into a couple of related, but distinct issues:

1. Technical: Are the applications generated by Flash 5 as good, or at least "close enough" to being as good as those created using Xcode? It's not a simple question, because even the best tools can be used to generated sloppy, bloated executables, but it's still important to get a sense of whether the typical Flash-generated application is going to be comparable to the typical applications created using Xcode. It's an unanswered question. I'm sure Adobe is going to argue that their applications are just as good. I, and many other developers who have been working with the platform since it came out are going to be somewhat skeptical of that claim. How likely is it that they've been able to add an extra layer of complexity, including a garbage collector, and not incur any additional overhead? Probably nots possible, but the question then becomes: just how much overhead does it add? Is it enough to really matter?
   
   This is just an educated guess, but with first and second generation iPhones, I'm going to bet it will matter, and the simplistic nature of the Flash apps already on the App Store does little to dispel that notion. If the apps were really "just as good", they would have put a sample up that really showcased what they could do rather than just showcasing the fact that they could do something. If they were just-as-good, Adobe would have wowed us with a 3D game or something else that really pushed the limits of the iPhone's hardware. Even on my 3Gs, I thought the Flash iPhone apps all felt a little sluggish compared to what I'm accustomed to.
   
   Unfortunately, we have no objective data on this point right now. We have no way to take an app generated by Flash CS 5 and the same or similar app ported to Objective-C and compare them in meaningful ways. If someone would be willing to compile a version of a iPhone apps with debug symbols on, I might be able to get some hard data using the performance tools. As things stand, though, we have no objective data, and without that, it's all conjecture. We could argue forever, and neither side will concede anything.
   
   In my mind, there's enough reason to believe the apps won't be "as good" technically speaking, that it's fair to put the burden of proof on Adobe. They should prove that these Flash-generated apps are capable of being just as good or at least identify how close they are and support that claim with verifiable data. We already know the file size of the Flash-generated apps are considerably larger than comparable Objective-C apps created with Xcode. Subjective evidence seems to point to the apps also being slower. I'd like to see some hard evidence so we can judge them more fairly, but until some exists, I think it's a fair assumption that these Flash-generated apps are not as good and the differential is enough to matter. If somebody wants to prove me wrong, I'd be thrilled to see some hard evidence so we can discuss this intelligently. If I am wrong, believe me, I will be suitably impressed by the engineering that went into it.



2. Logistical: There is a very large installed based of Flash developers and a large amount of existing content. The popularity of the App Store has already caused significant growing pains. The team responsible for doing App Reviews is overworked as it is and are doing a thankless job. Allowing existing Flash developers and existing Flash content to flood into the App Store almost overnight is bound to cause significant issues for the App Store infrastructure and the people responsible for maintaining it. That, in turn, will cause problems for customers and developers alike.



3. Equitable: Related to the previous point, this flood of new content literally overnight is going to have a significant and detrimental impact on existing developers. You know, the ones who invested the time and effort to learn the platform and native tools? Those developers who will be penalized for that effort by having longer wait times and a lot more apps to compete with for attention. It's unlikely that the size of the App Store pie is going to increase measurably as a result of the influx, so an influx of new apps is inevitably going to dilute the chances for any single application to make a profit. It hardly seems very fair that the people who actually put effort into learning the platform and who have dedicated themselves to the platform won't be able to reap the benefits of that hard work without putting in addition effort to figure out how to get noticed above a sea of quickly-ported Flash games. Contrariwise, it hardly seems equitable to reward those who couldn't be bothered to learn the native tools.



4. Political: Adobe and Apple long ago stopped being BFFs and have been in a somewhat contentious coopetition situation for quite some time. Apple made it very clear to Adobe on a number of occasions that Flash was unwelcome on the iPhone, and yet Adobe has continued to push ahead with an amazing disregard for that, fueled by what seems like a tremendous sense of entitlement: a feeling that they have a right to benefit from Apple's success with the iPhone even though they did not contribute to it in any way. The part that really doesn't sit right with me is that Adobe isn't being up front with their developers about how risky the situation is for them as a result of the politics of their relationship with Apple. Adobe didn't create Flash CS5's iPhone functionality with Apple's consent or knowledge. In fact, they did it despite being specifically told Flash wasn't welcome. That's pretty fucking rude. It's like inviting a million people over to somebody else's already-jam-packed party. Throw your own fucking party, man. Seriously. And the whole sneaky way Adobe went about putting their apps into the App Store just reeks of dishonesty. It's ends-justify-the-means thinking and it really does not endear me to Adobe at all.



5. Legal: Despite Adobe's claims, I'm incredibly skeptical that they could have created the tools needed to generate iPhone apps without using Apple's toolchain at all given the amount of time it took them, unless they had access to the SDK and violated the legal agreement that they had to agree to to get access to it. Apparently, Adobe management is reassuring their employees internally that they respected the SDK agreement to "the best of their understanding" and are claiming that they didn't reverse engineer anything. I suppose it's possible, but it's improbable to the point of ridiculousness and I don't buy it.
   
   That Adobe could have figured out how to link to and use objects and functions from Apple's frameworks and figured out what to link to without some amount of reverse engineering just stretches credulity given how much of the iPhone's internals are not documented officially. Adobe's trying to make it sound like it was just a matter of compiling down to ARM assembly code and they didn't need to reverse engineer anything to do that. But it's not as simple as that. These generated applications have hooks into many of the existing Apple frameworks, including UIKit, OpenGL ES, Quartz, Core Animation, and Audio Queue Services. They are leveraging Apple functionality to do a fair chunk of the work, which means Adobe did more than just compile ActionScript to execute on the iPhone's ARM processor. A lot more.
   
   Even if they didn't reverse engineer anything, I would put forth that Flash 5's iPhone application generating ability IS a derivative work under section 2.5 of the iPhone SDK agreement. This would seem to be precisely the type of thing that Apple was trying to prohibit when they put that clause in the contract, so even if Adobe was completely above-board and did no reverse engineering, I think there's still grounds to allege they are in violation of the SDK agreement.
   
   At very least, I would say that there's sufficient grounds for Apple to bring legal proceedings if they choose to. They could very likely get an injunction to prevent Adobe from releasing the iPhone portion of Flash CS 5 if they wanted. I have no idea if Apple will go this route. I honestly doubt it's their first choice for dealing with the situation, but it's an option that's available to them and, therefore, is an added risk for Adobe and any Flash Devs who choose to use it.

It will be interesting to see how things unfold. The technical questions will eventually be known. At some point, we will be able to compare Objective-C and Flash-generated apps and get some concrete data about their size, efficiency, and performance. For the rest of it, the ball is really in Apple's court. Until Apple makes their next play, we don't really have an inkling of how this will all play out. Everything from Apple embracing Flash IPhone apps to an all-out war with Adobe is within the realm of possibility.

Posted in: Adobe,Flash,Rant