What is a Workstation? (What should GNU ...)

[David R. Stampf]

>From: Marshall Feldman <RLN101%URIACC.BITNET at mitvma.mit.edu>
>Subject:      What is a Workstation?  (What should GNU ...)
>To: Unix Wizards Discussion List <unix-wizards at brl.mil>
>Message-Id:  <8908170656.aa09688 at SMOKE.BRL.MIL>
>
>Maybe we can focus the discussion if we ask about some interesting specifics:
>
>Is a SUN 386i a workstation?  Why? or Why not?
>If I soup up a 386 PC (on an MCA or EISA bus?) when, if ever, would it be
>a workstation?

	A few years ago, Bill Joy described a workstation in terms of 6 M's. I
think that the description is still fairly valid.

	Mip processing power
	Million Pixels
	Megabyte network
	Megabye of memory (hah!)
	Mouse
	Multuser/tasking Operating System

	This was about 5 years ago, and you could multiply each of these by a
an appropriate power of 2.  If you use this as a measuring rod, then the PC's
and Mac's of the world don't cut it, but a 386i easily fits.  

	The M's aren't arbitrary either.  They measure the user interface, the
compute power and your ability to connect to other like powered machines.  Some
boxes with 386's are workstation and some aren't.

>What is an AT&T 7300? (No jokes or flames please!  If IBM had made the
>original UNIX-PC instead of its MSAWFUL 8088 machine, we wouldn't be having
>this discussion).
>

	I often wonder what the world would be like had the 8088 based PC been
delayed a few years - no stupid memory limitations, no trying to teach novice
programmers the difference between near and far pointers, a real OS (a lot more
like OS/2 then CPM) etc.  But 640K looked infinite back then...

	< dave