[11] in Layered Athena
comments, and random other comments
daemon@ATHENA.MIT.EDU (Joe Harrington)
Mon Nov 16 12:42:45 1992
Date: Mon, 16 Nov 92 12:41:55 -0500
From: jh@oobleck.mit.edu (Joe Harrington)
To: marc@MIT.EDU
Cc: layered-athena@MIT.EDU
In-Reply-To: Marc Horowitz's message of Sat, 14 Nov 92 04:49:03 EST <9211140949.AA21030@deathtongue.MIT.EDU>
Cc: jh@MIT.EDU
Reply-To: jh@MIT.EDU
actually, i think the consortium xterm didn't have it either, but
athena did. the one that affects me now is that athena's xterms come
up one pixel wider than DEC's xterms. this sounds trivial but is very
annoying when windows come up missing an edge or sticking out further
than the other windows. problems in the past included a bug where on
an Athena VAXstation II/GPX, the mouse couldn't reach the bottom of
the screen because the mouse bitmap wouldn't go partially off-screen,
and the hot spot was in the middle of the bitmap. there were others.
i'm quite wary about athena's replacing any program that the vendor
supplies. even such utilities as the man program are different under
ultrix (ultrix won't accept MANPATH), and in some cases athena's
replacement is not entirely backward-compatible with everyone's OS.
for example, the professors in my lab will not use the machines in
their own athena cluster because when they type 'mail' they get an
error message and when they type 'rlogin' it waits a long time and
then tells them it's finally going to try to run the program they
thought they were running in the first place. yes, they are unwilling
to learn, but yes, it's their machine and they can and will say, "get
that thing off my machine" if it causes them even slight pain. many
other researchers in this department have the same attitude.
i'd like athena software and users' athena homedirs to be accessible
to all our users, but an option must exist where the default is for
changes to be invisible unless a user adds something to the path or
types the name of a program that wasn't in the vendor release. adding
directories of stuff is fine, adding new programs to /bin, /etc,
/usr/bin, and so on is even ok, but keep kerberized rlogin out of
those areas unless you call it something that someone else hasn't
already claimed, like krlogin.
perhaps one (easy?) thing to do that would make things more palatable
for researchers would be to athenify the few remaining standard tools
like mail (a LOT of people *really like* the berkeley mailer!), and to
add tools like ted t'so's getmail program (gets PO mail into
/usr/spool/mail) so that second- and fourth-party programs can be used
with athena's third-party system.
--jh--
