Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!pilchuck!pacer!davidb From: davidb@Pacer.UUCP (David Barts) Newsgroups: comp.sys.hp Subject: Re: pathalias Summary: "Here's your output from comsat, David." Message-ID: <297@zircon.UUCP> Date: 7 May 90 21:57:46 GMT References: <296@zircon.UUCP> <728@limbo.Intuitive.Com> Organization: Pacer Corp., Bothell, WA Lines: 85 In article <728@limbo.Intuitive.Com>, taylor@limbo.Intuitive.Com (Dave Taylor) writes: > Earlier in this group, David Barts commented: > > > When is HP-UX going to get a halfway decent version of pathalias > > (one that understands angle brackets and the FAST symbolic cost)? > [ justifications for HP not supporting pathalias edited out] . . . Well, they *do* supply it, and I think it's pretty dumb to send out a version of pathalias that can't correctly parse the UUCP maps currently being sent out over Usenet (especially when the most recent pathalias compiles with no problems on HP-UX). Kinda like shipping a free EBCDIC terminal with each HP-UX system and saying "Well, silly, *of course* you can't use it on an ASCII system." :-) > : > : > Perhaps fundamentally the problem you face is that you need to > improve your own UUCP connectivity? Hardly. I haven't had 10-minute transcontinential E-mail, but I can send messages to the East Coast and have them arrive in 20-25 minutes (which is the best UUCP connections I've ever had). > ... [ discussion related to docs for `X' being under `xserver' > and DT's opinion that HP docs are still the best ] ... Well, Solbourne (and presumably Sun, too) have beat HP on the `X' manual page -- by a little. The OS/MP X docs. have an X(1) which is the same as HP's X(1), but the server program (X) is under X(8C) instead of Xserver(1). It's still slightly sleazy, but at least I can find what I want on the second try. What would probably be best is to place all X windows commands in a new section of the manual (section 9? 1x?), then what was X(9) could be made into intro(9) and X(9) would document the server. This would be consistent with the rest of the manual. On a related note, why does HP place the maintenance commands and daemons in section 1m while BSD and Sun use section 8? Is Section 8 a BSD-ism, or did it exist in the V7 UNIX docs? If section 8 existed in the V7 manual, then is 1m a Sys{3,5}-ism or an HP-ism, and what was the rationale behind changing 8 to 1m? Which brings us to the reason for the strange "Summary" line. Back in my early days with UNIX in college (1982 or so) I was snooping about the /etc directory (VAX running BSD 3.9 or 4.something) and hit upon an interesting little executable called comsat. Neither the man pages not "man comsat" turned up any information, so since it is world-executable, I try running it to see what happens. Nothing... Hmmm, let's try feeding it the name of a text file. Still nothing? How about comsat -v (perhaps a verbose flag?)... Five minutes later I give up trying to figure out what comsat does (it is Friday afternoon). Monday morning, I walk into the terminal room and the sysop comes out and loudly dumps a 5" thick pile of fanfold on my desk, saying "Here's your output from comsat, David." 1000's of errors because the dozen or so comsat daemons with my UID didn't have enough permission to do their work. The (300 baud!) printing console had long since run out paper and worn through the printer ribbon. Turns out comsat was documented briefly in a paragraph at the end of the biff(1) page. So the moral of the story is that more documentation may save paper in the short run, but it costs in the long run. However, less documentation can make it easier to tell who's snooping about in the system files! BTW, I can't claim credit for it, but the next release of BSD did come with a separate comsat(8) page! But I digress. Probably the one thing that would improve HP-UX documentation most of all would be to have all the commands for each section in one place instead of scattered amongst the "HP-UX Reference", "Networking Reference", and "Using the X Window System" volumes. I really enjoy the bound (as opposed to loose-leaf) "HP-UX Reference" volumes, but I'd be willing to go back to loose-leaf tomorrow if HP would ship binders of a size (and the correct Permuted Index) so that I could assemble a documentation set with all of section N in the same place. Despite my criticisms, I'd have to say HP does a good job with documentation (but finding what you want can be a pain). Definitely better than the unspeakable barbarisms that have been committed to the Xenix docs! -- David Barts Pacer Corporation davidb@pacer.uucp ...!uunet!pilchuck!pacer!davidb