🚀 go-pugleaf

In article <108@oaxaca.nsc.com> ricardo@parns.nsc.com (Dick Smith) writes:
>Tait Cyrus, in his previous posting, brought up an interesting subject
>on which no one has commented.  Which is better, individual workstations
>for all users or X-terminals tied to a large server.  
>
>Is anyone else evaluating this alternative?
>
>Dick Smith                ricardo@parns.nsc.com
>National Semiconductor

It has been one of my tasks to evaluate different types of
X-terminals for usage in a VLSI design environment.  One of the
biggest issues that has come up is the amount of additional
network traffic that is generated by the X-terminals.  If the
programs that are being run use object oriented graphics there
is less network traffic than if bitmaps are sent to the
terminal but programs like Magic from Berkeley run some where
around 4-5x slower due to screen repaint times.  If you are
running xterms then a single X-terminal generates less traffic
than a diskless Sun workstation.  I don't have any other
vendors workstations running diskless on ethernet so I can't
really compare the two.  Also the estimated life expectancy of
the X-terminals in general is about 2-3 years.  At that time
almost all computers will have the ability to run the X-Window
protocol.  This ranges from Crays to IBM/MAC/Clones.  It is my
opinion after evaluating the options that the X-terminals will
be good for software development and debugging but I would not
want to run a graphics intensive application in an environment
where there were multiple X-terminals.

Mike Markley		markley%celece@ucsd.edu
UCSD Dept of Electrical and Computer Engineering
Apollo Systems Support.