Q: lseek returns long or int? (BSD 4

[Henry Spencer]

> Nobody (except for maybe Bill Joy) claims 4.2BSD to be PERFECT.

I've heard claims that sounded much like this from various Berkeleyphiles.
Maybe not perfect, but "it's so much better than everything else".  Nonsense.
Somewhat better, *maybe*, but hardly Nirvana.

> What do you
> EXPECT from a RESEARCH environment, a polished product?  By Grid, you must
> not have ever dealt with V6 - talk about rough edges!

An ARPA contract to improve Unix doesn't sound to me like research, it
sounds like software production.  (Yes, I realize that Berkeley got its
fingers burned and has sworn off this sort of thing now.)

As for V6...  He who never had to use V5 should not complain about V6.
I worked with V6 quite a bit, and didn't find it all *that* bad.
(Actually, V5 wasn't all that bad either, but V6 was quite definitely
an improvement.)

> I challange you to
> essentially REWRITE UNIX and NOT introduce bugs.  4.2 is the first extensive
> rewrite of UNIX for DEC processors ... so I can say I'm suprised.

What is not obvious is why Unix *needed* rewriting for Dec processors.
Oh, some parts of it, certainly, but most of it worked just fine as it was.
My biggest objection to x.yBSD is not the way they've fixed things that
needed fixing, but the way they've "fixed" things that *didn't* need it.
Of course new work introduces bugs; that's why people with sense keep
their hands *off* software that works fine and doesn't need fixing.  But
not Berkeley...

> I get tired of people slamming 4.2 - it's not perfect, but I'll take it,
> bugs and all, over all the other versions floating around.

I'll take my V7, thanks.  At least it didn't dump core all over my root
filesystem the first time I booted it.
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,linus,decvax}!utzoo!henry