On Wed, 11 Oct 2000, Joerg Schilling wrote:
>>>>>-rw------- 1 jmunsin infpersonal 5 Oct 10 12:42 dir
>>>>>-rw------- 1 jmunsin infpersonal 5 Oct 10 12:42 file
>>>
>>>>Perhaps it is time to upgrade. A fine young Finnish cs student
>>>>once created an operating system around your neck of the woods
>>>>that is quite portable. ;o)
>>>
>>>But obviously buggy. Did you report that bug to the Linux kernel team?
>
>>Actually, it is a feature. A useful one too. Time to update the
>>standards I guess.. Or will UNIX proper be around long enough
>
>There has been a long discussion on the UNIX 2000 mailing list.
>
>Just imagine, that some time ago, it was OK to open("", ...)
>to get the current directory. This is no longer allowed, you need to
>use ".".
If you think that is bad try this under VMS:
set default [.somenewdir]
create /directory [-]
Or something like that.. You can go into a non-existant dir
and then create it. It is braindead IMHO.
>If you would allow to open("bla/", ...) this would implicitly a open
>of "" in dir bla and this is the reason, why dir/ has no special meaning.
>For reasons of orthogonality, you need to open("bla/.", ...)
Just semantics IMHO. Linux isn't UNIX. It follows the standards
fairly closely, but not 100%. Where someone has deemed a
standard as braindead, it has been changed. Other areas -
standards don't define strictly and so are open to implementation
interpretation as well. I don't know (or care) about the 'adsf/'
semantic being in a standard or not myself, it works everywhere
I've used it, and I've seen it hard coded like that all over
hells creation so even if it breaks on every other OS out there,
it is a defacto standard in Linux, and UNIX has its own braindead
issues that are "standards" that technically could be changed for
a better OS, but aren't due to legacy and whatnot. Linux changes
over time to allow things in that make sense, but following UNIX
100% is not a design goal by any way shape or form, no more than
UNIX following Linux should be. Standards exist for good
reasons, but that doesn't mean that there aren't bad standards
too. Perhaps UNIX implementations, in an effort to follow Linux
standards, will upgrade to Linux semantics some day...
Doesn't really matter though.. Nobody's going to die either
way..
>>and in large enough capacity to matter now? ;o) SCO is no
>>longer with us... who's next the penguin asks... ;o)
>
>Let's see what happens with the Solaris source. Note that neither
>Sun nor SCO or it's new owner are allowed to make Solaris or SCO UNIX
>open source. Both in a joined action could do.
Unless they GPL it, or BSD it, it won't go too far I don't
believe. If they do GPL/BSD licence it, then that is great, and
they can benefit from existing Linux code, as well as Linux
benefitting from UNIX code. I'm all for it if it happens. I
just don't see it happening anytime soon. I think it will be an
11th hour decision that might come too late to make a difference.
Either way, the more choices we have, the better, and the more
open source we have the better choices we will have. When all is
open in this respect, the most used software will be the
"defacto" standard.
There are lots of changes yet to come, both in Linux, and in
UNIX. It's all for the good IMHO, even if I disagree with some
of the changes.
Take care!
TTYL
----------------------------------------------------------------------
Mike A. Harris - Linux advocate - Open source advocate
Computer Consultant - Capslock Consulting
Copyright 2000 all rights reserved
----------------------------------------------------------------------
"If it isn't source, it isn't software." -- NASA
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
