home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.sysv386
- Path: sparky!uunet!uunet.ca!geac!torsqnt!uunorth!anton
- From: anton@analsyn.gts.org (Anton J Aylward)
- Subject: Re: Dell SVR4 and Postscript ANSWER (from the NET of course)
- Message-ID: <1992Jul25.185215.9587@analsyn.gts.org>
- Organization: ASCI: UNIX Database and Communications
- X-Newsreader: Tin 1.1 PL3
- References: <1992Jul24.051353.3445@ttank.ttank.com>
- Distribution: na
- Date: Sat, 25 Jul 1992 18:52:15 GMT
- Lines: 46
-
- karl@ttank.ttank.com (Karl Bunch) writes:
- : .....
- :
- : (taliking about support calls)
- : It's important to make the whole situation a "team" event. Be helpful
- : as much as you ask for help. If you have 10 years of Unix experiance
- : spell everything out to the n-th detail so the other people understand
- : the entire picture. STOP WORKING FROM US -VS- THEM. Or else, you'll
- : get angry people wanting to find a way to PIN THE BLAME on you.
-
- In principle I agree. In practive I cry and tear my hair out.
- Two examples in point. I used to be a kernel hacker, I still deal with
- a number of "source" sites....
-
- Company 1, a TLA computer manufacturer, moved into the UNIX work with a
- vertically standing machine had a floppy disk driver that was f****d u*
- beyond beleif in rel 1.0. I called support, explained the phenomena and
- how the problem could be reproduced. Layered support; I got calls back
- to clarify and was handed up the tree. Eventually I got someone who
- could actually deal with source. I described the problem, told him
- where in fdstrategy() the problem was, told him what the code as written
- was doing, told him what it should be doing. He became verbally
- abusive, demanding that I treat him like a "professional".
-
- Company 2, deciding that UNIX is more profitable than oil, got involved.
- They have a setup script which uses the "file" command. This hangs when
- run on a tty line that has been exclusively opened. Again support call
- was 'resistive' because I could and did _exactly_ describe the phenomena
- and allow the support engineer to reproduce it. I explained how 'file'
- works - stat, decide if device or file, if file read, look for magic
- number, check in /usr/lib/gile/magic, if not then read and see it it
- looks like shell, c, fortran etc. I suggested that because the previous
- revision of the system worked - client still had some machines running
- it - that soemting had been changed, perhaps the order of the test.
- Abuse, core of which was "how can you possibly know that - you don't
- have acces to our sources!". Tell me guys, Middleton's PD version of
- withstanding, how many ways are there to implement 'file'?
-
- Yes, I agree with Karl, I've worked 'support' and would like as much
- detail, b ut some support desk seat warmers don't see it that way.
- Sigh!
-
- (BTW: that machine does have /usr/lib/file/magic, not /etc/magic.
- so much for standards!)
- --
- Anton J Aylward
-