att & osf

[Henry Spencer]

In article <3165 at homxc.UUCP> dwc at homxc.UUCP (Malaclypse the Elder) writes:
>...  anyway, judging from your last sentence,
>it seems that you are not sure whether newer system v utilities dereference
>NULL pointers or not.  if that is the case, how can you make a statement
>that at&t has done diddly to improve portability over the past several
>years? 

I have heard rumors that AT&T is cleaning up its act on this issue; in any
case, they have to now that they're in bed with Sun.  The statement stands:
over the past several years, AT&T has ignored all portability bugs except
those that cause trouble on machines that AT&T sells or uses.  The major,
possibly the only, reason that AT&T is now paying attention to the matter
of NULL pointers is that Sun sells machines that can't do *NULL.

>if it is true, as you claim, that porting unix utilities is difficult,
>then making them portable would be at least as difficult.  and if you are not
>sure whether at&t has done this difficult task, then how can you so callously
>be making such harsh judgments?

I didn't say porting the utilities was difficult; I said it was more
difficult than it had to be, because of AT&T's carelessness.  Getting the
*NULL bugs out of a large program can be a hassle, but it's not a major
nightmare.  Better yet, of course, is not to put them *in* in the first
place!  It's not that big a deal to fix things; Sun did it, back when Sun
was still a pretty small company.

If you want a really ridiculous example of AT&T's unconcern for real
portability, consider the SVVS.  But, you say, surely this is evidence
of AT&T's concern for portability:  a way of testing a new implementation
for conformance with AT&T's standard.  Except that a little bird tells
me that *THE SVVS ITSELF* has NULL-pointer problems!!!

>you may resent the fact that the unix system does in fact belong to at&t.

I can't say that I'm delighted that it belongs to a company that has done
such a lousy job of getting its act together, but "resent" isn't quite the
right word.  "Regret" is more like it.

>it is extremely expedient and self serving (not to mention a form of denial)
>to assert that the only entity that can claim ownership rights to the
>unix system is a bell system that is no longer existent...

Ah, but I didn't say that.  What I said was that the credit for creating
Unix went to the no-longer-existing Bell System, and that AT&T's right to
preen itself as the creator of Unix was doubtful.  It inarguably is the
owner.

> it is childish
>to take your frustrations out on a hard working development organization...

"Do not confuse effort with results."  "Work smart, not hard."  

>that in reality has done much good work over the past several years.

Change that to "some good work" and I won't dispute it.
-- 
Intel CPUs are not defective,  |     Henry Spencer at U of Toronto Zoology
they just act that way.        | uunet!attcan!utzoo!henry henry at zoo.toronto.edu