Saturday, April 5, 2008

Android Intended to Drive Fragmentation?

According to Sanjay Jha, COO of Qualcomm's chipset division, as quoted in this article in The Register, it was Google's goal from the outset to create, not reduce, fragmentation in the mobile software space with its introduction of Android. Qualcomm is, of course, a member of the Open Handset Alliance.

"Google wants fragmentation in the industry," according to Jha.

Why? Not too hard to figure out.

Google's Robert Love presented at GUADEC in Birmingham last year on the notion of a web-based desktop, where all your data lived on servers out "in the cloud" and was accessed pretty strictly via web browsers and web-based applications. This idea raised lots of criticisms, notably on the grounds of security and privacy (an area where Google's record is less than sterling), as well as accessibility. Love had no real response to the former criticisms other than "Trust us", and suggested that Google Gears could be used to address the latter, which kind of begs the question of why one should base everything on the web in the first place.

Google's shown that an area where they can succeed is in providing cross-platform tools for their properties such as Google Maps and the like, via precisely the sorts of web-based mechanisms that Love was proposing at GUADEC.

Why does Google want fragmentation in the industry? If the easiest way to make mobile applications is by falling back to the web, then that's a win for Google: they can cover a broad swath of devices while keeping platform-based competitors isolated to their own individual islands. This approach has the cynical side effect of marginalizing the work of the real open source community (y' know, the one whose source is actually open...?) with spurious claims that it "wasn't good enough" for Google's needs.

Google's approach to the mobile space has been deeply cynical from the outset: it's an open source project whose source isn't open, it's based on standards (live the Java language) while simultaneously ignoring the communities around those standards (like JCP), and it seems as though it's explicitly intended to hamper the development of a core mainstream Linux-based mobile stack by distracting the attention of potential developers.

I dunno, seems reasonably evil to me.