To: lkj who wrote (6432 ) 10/1/1999 2:41:00 PM From: Prognosticator Respond to of 10309
Hi lkj: thanks for your summary of ESC: I didn't get a chance to see the Live Code from Applied Microsystems, I had forgotten I wanted to see it, and only having 3 hours at the floor-show just wasn't long enough. I think it does require special hardware, as you suggest. Regarding tools, there are developers who think that GDB doesn't cut it, and those who find the Tornado suite cumbersome to use. Sadly, one of the best mechanisms for debugging these systems is a variant of the printf (via logtask so as to keep the intrusiveness down). Postmortem debug, i.e. crashing a task and then being able to connect gdb to see what was happening at the time is really useful, as is post-mortem when the system hangs (reboot and examine core). These aren't needed all the time, nor by most developers in my experience. WindView is useful when you're trying to tune your system, or track down a nasty priority inversion situation. Sun were showing micro devices based on KVM (one was a cellphone that's currently in use in Japan), and PersonalJava (a credit card terminal, where I rapidly purchased $800 of women's clothing. My wife will want to know about that one (thankfully it wasn't a real transaction). It showed the credit-card reader integrated with a neat LCD 3x4 inch screen, and Java ran snappily on it. What you saw at HP was CHAI: a non Sun-compliant VM that HP is pushing, and Microsoft is pulling, in an attempt to fragment the Java market place. I'm hoping that Pat Sueltz's appointment at Sun will lead to a resolution betweren HP and Sun, and leave Microsoft where they deserve to be. Anyway, to get to the point about RTLinux/Embedded Linux: what I was trying to find out with WIND was if they could see an opportunity to extend real-time Linux, with VxWorks, to get the best of both worlds. They didn't have an opinion. Which was a little disappointing. Next time they're asked, though, I'm sure they will have one. P.