[13790] in Athena Bugs
problems with infoagents
daemon@ATHENA.MIT.EDU (Fool)
Wed Aug 30 05:32:00 1995
Date: Wed, 30 Aug 1995 05:31:50 -0400
From: Fool <yandros@MIT.EDU>
To: bugs@MIT.EDU, cwis-dev@MIT.EDU
You guys really should get a mailing list for this. I'll add you to
infoagents if you like...
Anyway, there are a number of little problems with the infoagents
locker.
o /mit/infoagents/cat is probably a mistake
o /mit/infoagents/arch/share/bin doesn't fit in with the `common
practice' of new lockers; especially since you do have the more
`common' /mit/infoagents/share directory already, which is the
`right' way to do it.
o your `moved.sh' script generator is broken; it's added scripts to
replace binaries that never existed. In any case, you should
probably dump them all relatively soon; they're not used often
enough for the message to be necessary for very long.
o users on sun workstations have reported getting warnings from
netscape about references to
`/mit/infoagents/lib/NetScape/XKeySymDB', which makes sense, since
the directory is named `Netscape', not `NetScape'. However, I can't
find the reference to `NetScape' in the locker. Perhaps this was
already fixed and somehow I'm not noticing?
o I'd suggest removing the RT support altogether.
o You should be able to drop the inbsdbin symlink. nothing's using it
yet, and you'll be better off not starting it now and feeling like
you should keep it forever.
o your man directory doesn't have a whatis file. You have to run
catman on a sun AND a non-sun, not just on a sun. (yeah, this bites.
I've been meaning to do something about it, but...)
o for most of the browsers, the help files should be copied into the
locker rather than referenced over the network if possible (help in
lynx shouldn't send you to Kansas :-)
o another thing that I never finished with infoagents and yet which I
still insist is a good idea:
the default startup page for each browser should be a
browser-dependant file in the infoagents locker. This page should
contain announcements and information about the browser as well as
links to the places one might `normally' start a browser. This
also means that browsers don't take 2 minutes to time out on
startup if a particular server happens to go down. (honestly, I
just kept thinking of good things about this idea and I never
really came up with any bad effect...)
and a nit..
o /mit/infoagents/doc is really empty. There are a number of
documents that would be good to add there, such as the lynx user's
guide (should be in the source tree).
