Over on O’Reilly Radar there’s an interesting post about what it means to be an open platform (in the context of iPhone vs Treo). Some key points (which I think I agree with):
If you think about the original PC architecture, the openness was via the expansion ports and the commodity nature of the components used on the motherboard. It engendered whole new industries for adding functionality (e.g. video and sound cards) and scaring the platform provider (networking – who needs mainframes and minis?). But we’re all pretty much entrenched in that PC architecture. IBM of course lost control of the market to the clones and their attempt to close up the architecture failed miserably (MCA – micro channel architecture in the PS/2). But they did reap a lot of licence fees from the clone makers during that time (via patents on the technology) and I suspect they would still be doing so if they hadn’t tried to close the market up, thereby alienating the clone makers and forcing them to devise new alternate architectures.
Open platforms can be scary for the provider but if you hold your nerve, as we are with our open platform at Talis, then the rewards can be immense.
Those PC architecture parallels got me thinking about how the platform would enable the construction of a system by simply using building blocks from numerous vendors. Take that parallel one step further and you could expect to see providers specializing in the building blocks that they do best. Then I heard Tom Peters ranting “focus on what you do best and do it better than anyone else”