W(h)ither Open Standards for Mobile Linux?


One area of open source that I struggle with is that of mobile Linux: there are so many competing and intersecting initiatives that it is hard to tell who is doing what – and who is winning. Here are some fascinating thoughts from someone deep within that world, Jason Whitmire. He's General Manager of Mobile Solutions at Wind River – once one of embedded Linux's fiercest foes, and now one of fondest fans:

Limo, and in a slightly different way Android, have killed the standards-based approach to open source development in mobile. In the Linux world, creating an esoteric, theoretical application standard not based on market-driven code requires too much speculative investment without any clear mitigation of ROI risk for anyone to take up anymore. Indeed, the days of a bunch of representative techies flying to exotic locales to dream up the theoretical perfect system are over. It’s just too expensive to completely retool an entire stack without a known intrinsic return.

Because of the concentration in the mobile market (83% of handsets manufactured by five companies), when market leaders have invested in a stack, it is a standard whether certified by an arcane standards body or not. The Open Handset Alliance has created a de facto market standard not because a group of market leaders have adopted the standard, but because of Google’s overall singular market weight. The effect is the same (Trolltech and OpenMoko did the same thing that OHA did, and took it one step further by actually building a phone, but no one came running to embrace their reference designs because they lacked the market weight that Google has).

This is not to be confused with the need for technical standards that dictate interoperability, like GSM/WCDMA, TCP/IP, and WiFi (These were all the standards that LiPS cited as analogous reasons why there needs to be a theoretical application standard for mobile). These standards determine how unlike devices interoperate, which will always be needed. But within a device, the need for a theoretical standard is no longer valid. Developers will flock to what the market broadly supports, whether its because most of the Big Five support it or because of a sea-change in mobile computing brought about by an adjacent industry.

Provocative stuff.