Newsgroups: comp.os.minix
Path: utzoo!henry
From: henry@utzoo.uucp (Henry Spencer)
Subject: Re: Job Control in POSIX compatible Minix ?
Message-ID: <1989Oct8.031440.13499@utzoo.uucp>
Organization: U of Toronto Zoology
References: <5490006@eecs.nwu.edu> <3546@ast.cs.vu.nl> <3555@solo2.cs.vu.nl>
Date: Sun, 8 Oct 89 03:14:40 GMT

In article <3555@solo2.cs.vu.nl> maart@cs.vu.nl (Maarten Litmaath) writes:
>I strongly disagree.  The next included article I found on comp.unix.wizards
>just today.  I think Henry Spencer will have something to say about it, but
>this article makes clear the advantages of (properly implemented) job control.

Well, I don't read wizards any more (no time), so I'd missed that one.
However:

	1. Comparing a good implementation of a bad idea against a bad
		implementation of a good one often gives counterintuitive
		results.

	2. People confuse the awesome mess of job control with the simple
		idea of being able to suspend a process.  The two are not
		identical; see my more recent posting to lang.c and wizards.

In any case, you've missed Andy's point.  He didn't say job control wasn't
useful.  He said that it was very complicated and spread messy tentacles
everywhere, which is quite true.  POSIX 1003.1 is fairly simple and easy
to understand if you just ignore everything involving job control.  If
you don't ignore job control, it makes your head ache pretty quickly.
The complexity it adds is REALLY MASSIVE.  Read the standard if you don't
believe us.  Excluding it from a teaching system is sensible, and since
job control is optional in POSIX, the system is still standard-conforming.

(Andy and I both clearly believe that the horrendous jump in complexity
is a symptom of something badly wrong with the concept, but that is
another argument entirely.)
-- 
A bit of tolerance is worth a  |     Henry Spencer at U of Toronto Zoology
megabyte of flaming.           | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
