Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!utgpu!water!watmath!clyde!rutgers!ames!lll-tis!ptsfa!hoptoad!academ!uhnix1!nuchat!steve From: steve@nuchat.UUCP Newsgroups: comp.unix.wizards,comp.os.misc Subject: Re: Mach, the new standard? Message-ID: <300@nuchat.UUCP> Date: Fri, 28-Aug-87 16:26:34 EDT Article-I.D.: nuchat.300 Posted: Fri Aug 28 16:26:34 1987 Date-Received: Sun, 30-Aug-87 20:10:48 EDT References: <1665@ncr-sd.SanDiego.NCR.COM> <8381@utzoo.UUCP> <797@Pescadero.ARPA> <26310@sun.uucp> Organization: Public Access - Houston, Tx Lines: 32 Keywords: Mach Xref: utgpu comp.unix.wizards:3705 comp.os.misc:111 Summary: quick comment about bell's code [This has drifted away from architecture, comp.arch deleted from newsgroups list] I wrote: > > From what I've seen of bell's code it is a net liability anyway. To which Guy replies: In article <26310@sun.uucp>, guy%gorodish@Sun.COM (Guy Harris) writes: > That's a pretty broad statement; how much of that code have you seen? Some of > it is good, some of it is bad, some of it is in-between, and some of it is a > mixture of these. [I've cut out a lot of good points about the difficulty of replacing the utilities even if you do replace the kernel, etc.] From the code I've seen, which has been limited to 2.9 BSD and one very important sysV utility, precicely those parts that are somewhat hard to write are badly written. There seems to be a "catch-22" operating here. If something is moderately hard to do, like memory management, file system checking, or parser generation, to name a few, the first attempt is almost certain to be of less-than-polished quality. But since the job is both important and non-trivial, that first working piece of code is pressed into service, where it becomes entrenched. It is then immune to re-write attempts, and lives forever. I had written about several examples, but decided to spare the net and deleted the lines. Anyone interrested is invited to send mail. Steve Nuchia {{soma,academ}!uhnix1,sun!housun}!nuchat!steve