[net.works] Workstations

Deutsch.pa@Xerox.ARPA (04/17/84)

The Sun Microsystems workstation, while expensive at approximately $17K
for a networked, diskless configuration, and also suffering until very
recently from production engineering problems, has a good Unix system
with substantial graphics support and a growing volume of third-party
software including some good EMACS- and Star-like editors.  Warren
Teitelman, the principal designer of the user interface capabilities of
Interlisp, just went to work for Sun a few months ago.

"The basic idea is to take a lot of the capabilities now available in
LISP systems and give them to programmers in other languages" was one of
the motivating factors behind the Xerox PARC Cedar project, which
intended to do the same thing to the Xerox Mesa language which is in the
Pascal/Modula family.  Approximately 5 calendar years and at least 50
man-years later, a great deal has been accomplished, but you still can't
edit your program and be ready to run it again in less than about 15
seconds on a machine that's at least 5-10 times as fast as a 68000....

I suggest you think carefully about how to divide capability between the
workstation and any mainframes.  Bitmap workstations that can really do
everything for a professional programmer are still running in the
$10-15K and up range, once you count in all the file servers, extra
memory, etc. etc.  An alternative I've suggested to people in the past
is a bitmap terminal per person (should be around $3K these days) and a
VAX for every 5-10 people.  The key to this alternative is to
deliberately load the mainframe very lightly: it still may be
cost-effective, and it can handle "peak" requirements better than a slew
of smaller machines, although its responsiveness will be less even.

harry@uw-atm.UUCP (05/04/86)

This article was originally sent just to net.works on May 1.  Since I
do not receive net.works, please E-mail responses to me.  If
they are interesting enough, I will post them to Usenet.

-----

I work for an Atmospheric Sciences Department.  We are looking at
different workstations for manipulation and display of graphical
data, and radar and satellite imagery.  If anyone has any words of
wisdom about the following workstations, I would be very grateful:

Apollo, Masscomp, Sun, Dec

Harry Edmon       (...uw-beaver!geops!uw-atm!harry or
(206) 543-0547     ...seismo!uw-70!geops!harry)
Department of Atmospheric Sciences, AK-40
University of Washington
Seattle, WA  98195