Symbolics terminal emulator works lousy

[mcvax!jungfrau!ceb@uunet.uu.net]

In message <12358988425.16.CMP.SLUG@R20.UTEXAS.EDU> you write, (among
other things):

  ". . .  We keep our machines busy enough hacking AI programs to waste
  their time doing what some stupid unix box can do.  Lookit, the
  terminal program works quite well.  If you really cannot get up and
  walk to another console for mail (or have it arranged within
  swivel-chair distance), then use the terminal program to get to some
  boring, vanilla computer for which mail is about the most
  sophisticated thing you can do."

While in principal I agree with you; your philosophy essentially
avoids something I once saw where there were more Lispers than Lisp
machines, so people had to "sign up" (like in grade school), and
certain would do this just to read mail (!), kicking (me) off, who
couldn't really do the work elsewhere.

However, I simply do not find it true that Symbolics terminal programs
"work well".  True, chaos works better than the TCP stuff, but our rep
told us that if we had any more problems with chaos on our Unix boxes,
we were on our own.  So we switched to TCP, which leaves so many dead
copies of deamons lying around, we're thinking of hourly cron shells
to solve the problem.

Both SUPDUP and TELNET have had the problem for a couple of releases
that they make mincemeat of your screen when you use cursor motion
commands.  If you don't know that ^L refreshes your screen in Emacs,
once you read your mail using it through a TELNET or SUPDUP window,
you sure learn about it fast, because every other key you type is ^L,
in order not to be "editting in the dark" a la TECO.

Certain things, like Unix on-line manual support, just can't (or
shouldn't) be replicated on the ().

To SLUGgers in general: Pointers, tips, and refutations are gratefully
accepted.