Subject: Re: rm
To: None <>
From: None <>
List: current-users
Date: 10/31/1994 06:25:01
>     'proper' in my book does not include blindly following every
>standard that is plunked in front of you, just so that you can be
>'correct'.  BSD has been around long enough and is in wide enough use
>that it is a standard unto itself.  If ANSI, POSIX, XPG3, SVID2(3?),
>SPEC1170, or whatever conflicts with traditional BSD behaviour, then
>serious consideration should be given to ignoring the 'standard' and
>using the BSD method.  Not to mention, that there are so many
>'official standards' to which a UNIX-like system is expected to
>adhere, that there are bound to be conflicts between the standards
>themselves.  When I program on a BSD system, I expect it to behave in
>a certain way, this includes haveing things like setre[ug]id doing
>something useful.  I know this isn't an overly popular opinion, but
>every time I make a post like this, I do get mail from people agreeing
>with me, so I can't be totally out to lunch.

I do agree to some extent, however, in this case, I think it is more a
matter of consistancy, not with the standard, but internal consistancy. It
was pointed out in an earlier post that rm(1) works one way, rmdir(1) works
another, both on the same type of data. SYSV once upon a time treated the
trailing '/' the same as BSD still does, but at some point some person or
people decided it made more sense. This does not necessarily mean BSD has
to do this, or even necessarily should. However, just because BSD has
always done something a certain thing a certain way, does not mean that
this should also be blindly followed forever, standard or no.

Mike B