Nvidia Says Android Needs a Version of WHQL
Google needs to set more boundaries before Android fragmentation spins way out of control.
During a panel discussion on Wednesday during GTC 2013 in San Jose, Neil Trevett, VP of Mobile Content at Nvidia and four other panelists painted a nightmarish scene of developing for the Android platform. Yet at the same time, Google's platform was described as the next Windows: one operating system allowing for an infinite number of hardware configurations.
As a consumer, it's a pain (putting it lightly) when expensive software refuses to run correctly on a Windows-based desktop or laptop. Heck, it's almost just as annoying when an Android app continuously freezes or crashes to the home screen. Now imagine what it's like on the other side of the code, creating this service and trying to get it running smoothly on an untold number of hardware configurations.
Unfortunately for Android developers, this platform is seemingly repeating the early days of Windows before Microsoft implemented its Windows Hardware Qualifications Labs system. It's the fragmentation of the 90s all over again, only the form factor ranges from pocket-sized to hand-held instead of bulky desktops and laptops (yet).
"I think it's fascinating that the shape of the mobile industry is turning into the similar shape the PC industry used to be," Trevett said. "You have Apple which is vertically integrated, no single point of control, both beautiful and lovely. And then you have another platform that lets a thousand companies innovate and invest and make money. And so you end up with something that is actually more pervasive because a thousand companies can do more than any one company no matter how incredible that one company is. But you inevitably end up with things not being so polished and a little bit more chaotic. But that chaotic platform, because it's harnessing so much more energy, will win the majority of the market share."
Indeed this is happening, he added, as Android is now on 75-percent of the smartphones. He then goes on to point out a possible solution to Android's fragmentation, indicating that Google needs to take notes from Microsoft's success in taming the PC market, and enforce boundaries on what you can do and still call it an Android device.
"I think one of [Microsoft's] biggest achievements was going some way to avoiding this kind of [Android] chaos," he said. "WHQL [he pronounced it as "wickel"] for those of you in the PC space know that [as] Windows [Hardware] Qualifications Labs and their rigorous testing and putting in a few more specifications. That was actually Microsoft's, in my opinion, biggest achievement because you can do a lot with a PC, but it's not quite as chaotic as Android could be if this goes on much further. So perhaps Google should do something like WHQL. There's a lesson to be learned there."
But Android will likely not take the WHQL route. Why? Because Google is out to make money on advertising, and that means reaching as many eyes as it can. That's why it developed an open-source mobile OS so that ODMs (and consumers) can take the platform and run. That's not saying Google isn't working to reduce fragmentation – it is to some degree, but you're only going to support free software so far. Meanwhile, Microsoft has a WHQL testing fee of $250 USD which seemingly goes against the cost savings principles of an open-source platform.
As covered in one of the previous sessions, Apple has seemingly stayed the same throughout the decades: sell an extremely small number of form factors with a finely-tuned, stable software platform. Microsoft has focused mostly on software and reaching as many customers as it can through a number of ODMs. Google is doing the same, yet it's not about selling additional proprietary software for that platform, but to sell additional ads. This may be why fragmentation will only continue to get worse.
The whole conversation about Android fragmentation stemmed from a question to the four panelists – SoftKinectic VP Tim Droz, Metaio SVP of OEM Business Development Graeme Finlayson, SmugMug CEO Don MacAskill and Sensor Platforms VP of Engineering Jim Steel – about what they would put in the next mobile SoC. All of them wanted sensors, but added that it would escalate the already daunting fragmentation problem.
"[Fragmentation] is a huge problem – there's not a lot of debate about that," the SmugMug CEO admitted. "It is from our point of view the single biggest reason why it took us so long to – and why we didn't initially target – build on Android. We are a photography app, and so the myriad of sensors for the cameras is a huge problem, and the myriad of different displays for providing pixel-perfect really well-designed interfaces to let you do the things we think you should be able to do to your photos is a big deal."
He's obviously talking about the numerous camera sensors and display sizes and resolutions provided on a multitude of Android smartphones and tablets. Compare those numbers to the iPhone, iPod Touch and iPad units, and you see why Android development can be a scary, scary thing.
"That's just talking about the hardware," he added. "The software is yet another issue. We've talked to some of our friends in the industry who have test labs of more than 700 devices that they test on, and we just can't go there. We're a small company and so we're forced basically to pick and choose."
Who should try to work on this problem? Hopefully Google is working it, he said. "We've been asking that for years now," he admitted. "[But] the problem seems to be getting worse so I don't know [the answer to the problem]. I wish we did."
The solution may be a WHQL model, or rather an AHQL, but whether that happens or not is up to Google itself. There was quite a lot of buzz during the conference about the X Phone series and the ability to customize a phone, and the idea seems really great from a consumer standpoint. But this custom-built form factor could ultimately make the fragmentation matter much worse if Google doesn't set some type of WHQL-like boundaries.