I want to emphasize that I think the relatively recent notion of EIDs are so important as to have altered my view of the outcome ubiquitous computing. A year ago I held the view that embedded systems would be dominated by large electronics consumer products companies, each banging out lots of intelligent devices adding to everyone's enjoyment of life. Consequently, I looked forward to WIND's run-time royalty stream becoming the mainstay of the business in the next Century. I thought that embedded systems programming would never become mainstream because the inherent hardware/software complexities necessitate the kind of multidisciplinary team that could best be fielded by organizations able to write-off high development costs over large production runs.
After thinking more about it, I now believe that, in addition to the development of smart consumer products, virtually every organization will begin first to augment their IT systems with EIDs, and then to develop their IT systems around EIDs. For this to happen, incorporating EIDs must become similar to integrating a database into an application. Mainstream programmers must become comfortable programming EIDs, and including them into applications, along the lines of how they currently build client-server applications.
Hardware engineers prove themselves everyday by meeting Moore's Law. Due to the relatively large cost of hardware development and production, their work tends to come first, laying the foundation for product development. Unique hardware components that will comprise EIDs will be in thousands; whereas unique software modules laid on top the hardware will measure in the millions or even billions. This means software becomes the bottleneck choking deployment of embedded systems, especially those hundreds of thousands of soon-to-be EID applications.
WIND's Tornado IDE will continue to facilitate development of systems deploying EIDs. The company has announced complete package solutions of needed components including providing help selecting and configuring appropriate hardware - a godsend for software-oriented designers. I suspect the Tornado environment will continue to expand, both natively and through partners, to provide programming tools that are more and removed from hardware specifics to enable mainstream programmers to pitch in. Just as Oracle hides all the details of data storage and retrieval and is generic across platforms, Tornado will shoulder an increasing responsibility to hide hardware details while covering the plethora of possible underlying hardware configurations.
What does it all this mean for WIND's revenues and earnings? Mostly, it means that WIND will not converge to just a royalty coupon-clipping company. WIND will always have growing product license sales along with growing engineering services, keeping royalties somewhat in balance. (Actually, Ron Abelmann has been saying this from the get go, but I have been slow to fully appreciate his reasoning, and secretly doubted him. Wrong.)
Royalties will be larger than I ever imagined, because not only will all the huge production runs happen as I always anticipated, but with EIDs, there will be many more embedded products using VxWorks in relatively small production runs than I would have thought possible. Don't forget, small production runs usually carry higher unit run-time fees.
Allen |
