🚀 go-pugleaf

Randy's message from DARCOM made the point that their
workstations and their big machines run "the same" software.
Then, he mentioned that the "send" command on his little machine
doesn't actually get the mail to the ARPAnet by itself but rather
hands it on to one of the big machines to do the job.  So,
really, the different machines don't have/need the same software
because they don't do the same jobs.

This highlights an interesting issue, that of the users' view of
where the capability resides.  Maybe what he is really saying is
that the users have just one set of commands and the machines
between them divvy up the work.  This is really what an
integrated user interface is all about.  The users know what
capabilities are available and how to request them but can and do
remain unaware of what computer actually does the job.

Large jobs (for example, formatting a 300 page document) ought to
be handled on large computers.  Doing it on a small computer, you
have two kinds of choices -- make the program very small (not
powerful enough for the job) or make the program very slow (tying
up your workstation for hours).  With a large computer lurking
nearby, your small machine can send the big job off for
appropriate processing.

Of course, having big machines in the basement costs money.  How
does DARCOM go about factoring in the big machine cost into the
$5-6K per station figure?  Was that hardware cost or a system
price that includes development costs?