FERGUSON@TMASL.EXXON.COM (07/01/89)
Thanks, fclim, for the extra comments. I haven't seen other replies
that you mentioned (like Darryl's), I think routing to bitnet from the
internet is less than straight.
Anyway, here's a pretty sweeping statement that will have less than
a major effect on anyone, but I thought I'd say it.
I think the world can do a whole lot better than UNIX.
I'm not saying that Aegis is the greatest either, but I don't see why
we're diving over the standards cliff just yet. PC vendors like
Commodore-Amiga and MacIntosh have shown a ton of functionality in
having icon-ized operating systems. Not only is a command-line
interpreter, but those who wish can run their applications without
knowing anything about the 'shell','cli', or whatever you call it.
UNIX loves TCP/IP, but TCP hasn't been truly accepted as the best
way to communicate.
UNIX loves Ethernet, but we all know that throughput is pretty lame
on Ethernet because of traffic, collisions, and only 10 Mbit/sec.
People will reply that UNIX is independent of TCP and Ethernet, but
I think they're pretty well stuck together right now.
I may be slightly uninformed about this, but isn't the ISO moving
toward something other than TCP/IP?
People clamor that they want their UNIX 'bug for bug' compatible. Why
should the world embrace a technology so dependent on mistakes?
Sendmail was way open for that guy from Cornell to demonstrate the
weakness in UNIX's inter-system mail software.
I haven't seen sr10, bsd4.3, or system 5.3, but I don't think one
level of revision is enough to create our Utopian operating system.
I think we can borrow a lot from our PC friends (as long as they
don't sue us for look and feel copyrights) without losing performance.
And I don't buy the argument that only fools will use a system
that's so simple a fool can use it.
If you bomb out a process, can Unix give you a traceback of what happened?
Should we let AT&T, OSF, and others with boatloads of money force this
on us, just so that they can stop spending so much on OS research and
tell us that they're getting the best OS that you could ever want?
Anyway, that's my two cents about UNIX, and it wasn't even a pro-Aegis
argument, and it wasn't meant to flame on anyone. I just think that
before we make standards, we have to know that it's a waste of time
developing this stuff any further. That means that OS revisions would
be unnecessary, and those bugs that people love so much wouldn't exist,
and there would be no more arguments over what's a bug, and what's a
feature.
Thanks for tuning in,
Scott Ferguson
Exxon Research & Engineering
Route 22 East, Room LD280
Annandale, NJ 08801
(201) 730-2339
ferguson@erevax.bitnetGBOPOLY1@NUSVM.BITNET (fclim) (07/01/89)
> From: FERGUSON%TMASL.EXXON.COM@CUNYVM.CUNY.EDU > Subject: UNIX and things > > If you bomb out a process, can Unix give you a traceback of what > happened? I'm going to open my mouth again and let all know I am a fool. I think (ie I'm *not* sure) that if Apollo allows the image of the process to be dumped (known as a core dump in the Unix community) whenever the process bombs out, then we could use dbx, adb, sdb, etc to debug the core. There should be a traceback somewhere in the core. fclim --- gbopoly1 % nusvm.bitnet @ cunyvm.cuny.edu computer centre singapore polytechnic dover road singapore 0513.