>Well, Allen et al, how about this new item on potential competition?
WIND is no more threatened by a no-name software company, in this case Integrated Chipware, than Microsoft might be from a new entrant in the desktop OS business. The article misstates and overstates every comparison with commercial RTOS's, and is a waste of ink and paper. As a matter of fact, if you show almost any hyperbole from the article to a representative from WIND, or INTS for that matter, they will think the statement is about their RTOS. Consider: That ability to tailor code for a specific product is a key differentiator, according to Hodges. The icWorkshop includes a real-time operating system, but its mainstay is its tool kit. That kit lets programmers write their own code or pull together different pieces of the RTOS, using only a minimal amount of code. That's what VxWorks allows, even encourages-at a granularity far beyond modules.
Integrated Chipware is not a revolutionary advance in RTOS state-of-the-art. It is yet another "no-royalty, you-hack" OS that is at most of passing interest to passersby. Imagine yourself as a manager of a group trying to get products to market. You finally built up enough nerve to inform your OS group of your recent conclusion that maintaining and extending various versions of an in-house OS is no longer justified. You are going commercial. After making that decision, the last thing you ever want to see in your organization is source code for an OS.
Let me try to explain this another way. Organizations facing competitive pressures to get product out the door absolutely must elevate their abstraction of the development process. Modern, successful development organizations leave the OS stuff to OS pros, the add-on stuff, like graphics, communications, file systems, etc. to add-on pros. They concentrate strictly on the value they add to the product--algorithms as application code.
If a developer touches the OS, by tinkering with the OS source code, they own it. That means they have to take full responsibility for maintaining, porting and upgrading the OS source code for the complete life cycle of the product. Why would anyone departing from the home-grown OS minefield willingly take on such an awesome responsibility?
Of course the royalty-free business model will be attractive to a few low-ball developers. But the chaotic world of embedded systems has been awash with royalty freebies from the outset without noticeably slowing the market leaders. There is nothing new or different in this regard about Integrated Chipware. If I offered you the "Super Windows" desktop OS for free, would you use it?
The fact is that a primary focus of attention at WIND is the practical accomplishment of virtually everything espoused in the article about Integrated Chipware, with the exception that WIND does not provide scalability down to the level of source code. The WIND kernel can be as small a 20K bytes, with the WiSP kernel much smaller yet. It can be tailored with granularity beyond the module level, by selecting portions of support libraries, as erroneously denied by the article. The BSP has long enabled the kind of porting claimed to be a new feature of Integrated Chipware's OS.
Integrated Chipware and numerous other no-royalty OS's are no threat to WIND. WIND's competitive energies are directed at other serious commercial RTOS vendors, like INTS, and Microsoft.
Allen |
