The embedded market demands a lot of its operating system. When this market needs something small, it has to be truly lean. When they need more features, such as networking, removable media support, and a GUI, developers would prefer to build on their established base, rather than starting from scratch with a different operating system. The field is also populated with a lot of folks who know how to do things for themselves, including getting up to their elbows in code. Finally, embedded devices are often consumer goods, and thereby designed for high volumes at lower profit margins than custom-built equipment. So the traditional royalties that developers have had to pay for both the tools and runtime licenses now become profits or can be invested back into R&D.
Some pretty big names seem to think that Linux fits the bill for the embedded market, and are throwing their weight behind the effort. (See the roster list at the Embedded Linux Consortium.) Does this mean that Linux will take over the embedded market? Some say that it's a foregone conclusion; others are cautiously optimistic. On the other hand, you won't find many nay-sayers. Linux has been available for x86, ARM, and Motorola 68k for a while nowgiving developers on these platforms an established toolbase and headstartand Linux has recently been ported to the Palm computing platform. To paraphrase James F. Ready in an interview from the premier issue of Embedded Linux Journal, you're going to see more Linux in the embedded market because it's the platform for which most engineers prefer to develop.
Regardless of whether the next "killer embedded app" is Linux-based, participation by the embedded community is good for Linux. Know-how for real-time Linux will increase, benefiting other Linux markets such as network appliances and routers. Linux will gain increased exposure as consumer products are sold as "based on Linux." And Linux will be ported quickly to new embedded architectures by the community at large because a) it will be worth an embedded Linux company's effort to assist in the port in order to use their existing toolset and codebase; and b) as more tools are shared and released with open licenses, more development time can be spent on building products, not tools, bringing more people to take a look at Linux, and so on
In conclusion, it would be unfounded to say that Linux is "losing ground" because it had a tough year in the stock market. I'm not sure that Linux is ever going to be the darling of Wall Street (or let itself be exploited by Wall Street, depending on how you look at it)it's not that type of operating system. To put it in perspective, we're talking about a piece of technology that lays its complete design on the table for all to use, and has gained popular acceptance without advertising, a budget, or an angle. (Would anyone like to revisit their Windows versus Macintosh choice in the late 1980s and early 1990s?) Linux came into popularity based on technical meritso I'm not sure that traditional industry indicators are applicable.