From: utzoo!decvax!harpo!floyd!vax135!ariel!houti!hogpc!houxm!ihnp4!ihuxx!ignatz
Newsgroups: net.misc
Title: Re: netiquette - (nf)
Article-I.D.: ihuxx.409
Posted: Fri Apr 29 14:16:34 1983
Received: Sat Apr 30 07:14:42 1983
References: <1961@uiucdcs.UUCP>

Pardonnez moi, but all of this one-notesfiles-using-PLATO-system-generates-X
thousand-articles-a-day-and-that's-more-than-all-of-USENET-so-there is bunk.

PLATO does *not* run on a minicomputer.  Correct me if I'm wrong (silly me...of
course...you guys wouldn't miss the chance!), but most machines running UNIX
are minis, with a few micros lurking in the background as plain uucp sites.
(Yes, you, Lauren!).  I haven't had the pleasure of using notesfiles, but I've
seen the interface in use by others, and yes, it's better.  But the thing is
a pig.  A solution that works on a CDC Cyber 7 (or whatever PLATO considers
its brain these days--it's still certainly a mainframe) will most likely not
be acceptable on a mini-based UNIX system; in fact, it isn't even so great on
an IBM 3033-based UNIX!

Just as an aside, I'm amazed and a little concerned at the number of young
grads (Huh? Gramps? Oh, well, I'm 29, and I've been out 'in the real world' for
about 7 years.  Yes, a recent grad is young.)(*my lumbago!*) who've never worked
on mainframes, and devoutly maintain that "anything a mainframe could do, can
be done by my micro" and "UNIX does it all".  This attitude is reflected in
the simple comparison of a PLATO-based system and a UNIX-based system on the
grounds of the volume of notes.   NO, Virginia, a mainframe and a mini don't
compare.  NO, Virginia, UNIX doesn't do it all.  It's dandy for what it was
designed to be, but that isn't everything to everybody; which is what most
mainframe OS try to be.  Granted, the code is often a nightmare, and bugs are
often as easy to track and zap as roaches in a Chicago tenement; but the
things do more, faster, than minis.  Look at simple old UNIX when ported to the
IBM; it's already grown warts and sacks to deal with design deficiencies that
surface when dealing with the greater resources of a mainframe.

I think mainframes as we know them are a dying lot; but not because a mini like
the VAX will do them in.  I think network technology and techniques, coupled
with increasingly faster hardware, will allow a network of minis to perform
as a mainframe with infinitely variable power.  But for now, a single mainframe
will almost always be able to outperform a single mini with respect to volume
of throughput, computing power, and speed. (Unless they point out an egregious
error, unique and exceptional counter-examples to the bit bucket.)  Please
consider software performance in light of the support environment, eh?

		"I ain't religious, but a human's gotta preach sometimes..."

		Dave Ihnat
		ihuxx!ignatz