Allen,
The the article link Daniel Willert posted last night started me thinking some more about the whole Linux/"open source" issue. A few months back I seem to recall your mentioning the Linux open source issue WRT WIND and then saying something like "I'll be providing some more detail on that later." I (and I'm sure others on the thread) would really value your focused analysis/feedback in this area. (If you've already posted such an analysis and I missed it, please forgive me.)
Would also appreciate any feedback you (or others on the thread) could provide on my "stream of consciousness" musings on this topic below.
My musings:
A major flaw I see in the "open source" argument is the alleged "vendor independence" it provides. In order to have true vendor independence, customers need to have not only an open source OS, but also open source tools and layered applications. Once customers become dependent on a single Linux vendor's tools and layered applications (or particular integration thereof), it seems to me that they're still "locked in" to that vendor--regardless of the fact that the OS is "open source." Of course, a customer could conceivably remain completely vendor-independent if they insisted on building their products only on "open source" tools. But then the questions that arise are: (1) Which open source tools/layered apps should I use?; (2) Which open source tools/layered apps will "win" and which will end up languishing without support? (3) How do I ensure that the tools I selected work together? The answer (according to Linux vendors) is "let us handle that." But then, that just leads to an additional question: Which Linux vendor should I choose? The answer to that question is (of course): "the one that's going to win in the marketplace and, therefore, won't go away in a few years leaving me with no support (or only informal support through the open source community)". But isn't this the same question that customers are also asking about the other (non-Linux) embedded OS vendors? If so, then what's the real benefit of going with on "open source" OS? The main "benefit" I see is that the customer does still own the source code and can still leverage the benefits that the open source community contributes. However, I would argue that that benefit is not as great as it seems because of the following unknowns for the customer: (1) Will the open source community continue to support all the components I've chosen for my product even after (in Steven Stolper's words) "the programmer graduates from college"? (2) Even though I have the source code, will I really want to use my precious resources to do anything with it?
All the above leads to what I think are perhaps the two most important questions for embedded OS customers: Which of these company has the best business model to support a stable, long-term business? Among the non-Linux embedded players, WIND is emerging as the clear winner. Among the Linux embedded players, the jury is still out on a market leader. So the choice really is between:
1) A market-leading proprietary OS vendor with a rapidly expanding value chain, high barriers to entry, and a revenue stream based on both royalties and services.
2) One of "a basket full" of Linux players whose R&D is performed largely by an informal "open source" community, but with low barriers to entry and a services-only based revenue stream (unless they put in proprietary value-add, thus violating the "open source" ethos which they so fervently tout). |
