D. Peschel said...
|
|What _I_ want to know is the relationship between OSF/1, a.k.a. Digital
|UNIX, and AIX. The manual entry for stanza(4) on our DEC system is the same
|as OSF/1 1.0, and it's SO badly written that only IBM could have done it.
|Plus stanzas are an IBM-ism anyway. I bet the entry is pretty much the same
|on AIX systems, too. (I just checked and our AIX doesn't have it. So I
|don't know -- it's just a guess.)
DEC first seriously attacked the UNIX market with Ultrix, which
was almost a pure BSD port. Over time they added a bunch of stuff
to it. They had a few really cool things, like they dxdiff, easily
the best GUI-based diff I've used.
Then along came OSF, a hodgepodge of UNIX vendors scared of the
idea of Sun and AT&T getting together and taking over the world.
So they put together OSF/1. This included people from a whole
mess of computer vendors, but DEC and IBM wre probably the biggest
contributors. The reference platform was a DECStation ?000 (I
forget which one).
By the time OSF/1 actually came out, it was obvious that it was
a fair example of design by committee, and the AT&T/Sun alliance
had pretty well faded away. So DEC, for who knows what reason,
oter than that they had the goods, decided to swap Ultrix out and
OSF/1 in. And everyone else went back to whatever they had been
doing before - AIX, HP/UX, whatever.
OSF actually did some cool stuff, but they had some real problems,
too. I was working at SecureWare, who was doing the security stuff
for OSF/1, and I remember new sandbox drops every day, and bugs
that had been fixed reappearing over and over in their original
form, and just general mayhem.
My favorite part, though, was when we built the kernel and X11
fully debugged, to track down a really nasty problem. Just booting
the OS ate up 70 or 80 MB of the 128MB in our DECStation, and
adding X11 on the hi-res screen ate up the rest of the RAM and a
fair bit of swap space. Then I tried to log in, which took almost
an hour exactly, beating the snot out of the disk drive, until it
finally came up - having used over 100MB of swap space. Starting
a Motif app ate the rest of our 256MB of swap space over the course
of two hours or so, at which time the system just rolled over and
stuck its little feet up in the air.
We never found the bug.
Three or four builds later, it was gone, somehow. Nobody at
OSF knew where it came from, where it went, or why. Oh, well.
My desktop machine at the time was the last of the Tektronix
68K-based systems running a BSD variant. It was slower than
their ill-fated 88K systems, but those ran SVR4, and mine was
fast enoug for me. It was a spiffy machine for the time. Of
course, I was just glad I wasn't running A/UX on really slow
Apples like most of my co-workers!
-Miles
Show replies by date
D. Peschel said...
|
|What _I_ want to know is the relationship between OSF/1, a.k.a. Digital
|UNIX, and AIX. The manual entry for stanza(4) on our DEC system is the same
|as OSF/1 1.0, and it's SO badly written that only IBM could have done it.
|Plus stanzas are an IBM-ism anyway. I bet the entry is pretty much the same
|on AIX systems, too. (I just checked and our AIX doesn't have it. So I
|don't know -- it's just a guess.)
I believe AIX was one of two (along Digital OSF/1) Unix varients
based on the OSF/1 code. IBM also contributed the file system
and LVM to OSF/1. DEC kicked in a bunch of their parts.
The Sun-AT&T deal faded away as AT&T decided to buy NCR, partner with
Pyramid and continue their attempt to chase the computer business
with badly conceived plans (like buying NCR).
Bill
--
" ... Overall we've found FreeBSD to excel in performace, stability,
technical support, and of course price. Two years after discovering
FreeBSD, we have yet to find a reason why we switch to anything else"
-David
Filo, Yahoo!
Bill Pechter said...
|
|I believe AIX was one of two (along Digital OSF/1) Unix varients
|based on the OSF/1 code. IBM also contributed the file system
|and LVM to OSF/1. DEC kicked in a bunch of their parts.
I don't doubt they rolled some of the changes into
AIX, but (a) they had already put quite a bit of AIX
into OSF/1, and (b) they certainly didn't start over
with OSF/1 as a new AIX base.
|The Sun-AT&T deal faded away as AT&T decided to buy NCR, partner with
|Pyramid and continue their attempt to chase the computer business
|with badly conceived plans (like buying NCR).
And also because Sun was freaked out over OSF/1, and because
their much vaunted UNIX to end all UNIX didn't pan out...