On Tue, 2002-09-10 at 15:02, Shawn wrote: > Specific ports to specific processors would mean more focus on what > makes each one tick, and how to "tweak" them into their own dynamic > realms. Sort of like how each variance of Unix is for the most part > specific to their platforms: HP-UX to PA-RISC, Solaris to Sparc, etc... > > To me, that would be the biggest overall part of it. The downside is > that it's more work on the developers as well as the maintainers. Well, this is similar to how Unix got really fragmented.. Different groups and companies would go make their own version. Much of what you're talking about ends up being a compiler problem rather than a kernel problem anyway. Sure, there are ways that rearranging data and procedures can help, but it would be tremendously painful to write code for each different processor, and each sub-version of each processor, etc. -- _ _ _ _ _ ___ _ _ _ ___ _ _ __ If you tell a joke in the / \/ \(_)| ' // ._\ / - \(_)/ ./| ' /(__ forest, and nobody laughs, \_||_/|_||_|_\\___/ \_-_/|_|\__\|_|_\ __) was it a joke? [ Mike Hicks | http://umn.edu/~hick0088/ | mailto:hick0088 at tc.umn.edu ] -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: This is a digitally signed message part Url : http://shadowknight.real-time.com/pipermail/tclug-list/attachments/20020910/8694b28e/attachment.pgp