uid=gid argument

[birkholz@midnight.com (Matt Birkholz)]

   Date: Wed, 16 Mar 94 21:46 PST
   From: bruce@beagle.pixar.com

   I'd like to call time on the uid=gid argument. I'm convinced this issue
   is not going to come to resolution through discussion here. Perhaps a
   demonstration would be more appropriate.

I _did_ call time on "the uid=gid argument".  (Again, the misnomer!  See
how these things spread?)  However, after a gazillion non-arguments against
it, I'm afraid to see the proposal defeated by "proof by repetition".  (I
hope I don't regret that as much as I regret mentioning "the principle of
least astonishment". :-)

If you really want to call time on the argument, I'd suggest making an
example of yourself.  Calling time on an argument and then slipping in a
couple extra jabs seems rather hypocritical.

As for the "extra jabs":

   The people who feel uid=gid is important are perfectly capable of
   building a package to implement it on top of a newly-installed release
   [...]

And vice versa, e.g. a SYSV compatibility package for the terminally
paranoid.  The issue here is what the default should be.  Telling us that
the default should be _your_ favorite because the rest of us can just go
write a script isn't exactly an argument.

Here _is_ an argument: if we make setgid directories the default, people
will notice that group permissions can actually be useful for project
directories.  If some people are scared silly by the 's' where there
"should" be an 'x', they can apply the "Dark Ages" patch and return to the
comfortable fold of the one true church.  At least opportunity bothered to
knock on their door.

   Please don't ask everyone to take it on faith that uid=gid is
   appropriate.

There's no faith needed.  It works.

I'd really like to know who started calling this the uid=gid debate.
I'm not even sure we're all talking about the same thing anymore.  Can we
call it "setgid directories" or "private groups" or "umask 002" or
_something_ sensible?

I'm having trouble imagining anyone with anything better to do worrying
about those 's's.  Have we all wrung our hands over the fact that the group
of /usr/man is some strange group named "man"?  It works, so we don't care.
(Or do we?  Am I going to regret mentioning this feature?  Will we also be
needing two man packages: a traditional, non-working package; and Daniel's
"21st century man" package?)

   And don't consider it overridingly important that you "save" every user
   by getting uid=gid into their system now.

Actually, we're just trying to save ourselves.

					      Put together a package that
   implements it. Show us how many users you get, and how they feel about
   it, and then you'll have the ammunition to show that everybody's system
   should be configured that way.

Fine.  I don't care.  The unwashed masses can remain unwashed.  Debian will
not be the first broken Unix system with which I've had to deal.

Matthew Birkholz
birkholz@martigny.ai.mit.edu