[8073] in Athena Bugs
Re: rsaix: info
daemon@ATHENA.MIT.EDU (Richard Basch)
Thu Aug 29 22:12:58 1991
Date: Thu, 29 Aug 91 22:13:18 -0400
To: ckclark@MIT.EDU
Cc: basch@MIT.EDU, bugs@MIT.EDU
In-Reply-To: Calvin Clark's message of Thu, 29 Aug 91 11:37:15 -0400,
From: "Richard Basch" <basch@MIT.EDU>
I should be more specific now that I've tried it again. When running
the tty based info client in an xterm, portions of the screen become an
unreadable mess. When I tried aixterm, it worked better, and I was able
to get some information out of it now that you've told me how to select
menus with CTRL-o. However, sometimes it gives warnings stating that it
can't bring up a graphic, which is the proper thing to do if it can't,
but some useful information is lost that way, especially if the
"graphic" is really just text which IBM, for some reason of dubious
merit, has decided needs to be displayed in PostScript. You say that
you avoid the tty-based client, and I can see why, but at least man
pages can be viewed without an X display.
I am not disputing that it is a bug that it did not display correctly
within an xterm.
It is not possible for me to
bring up the X based Info Explorer remotely on a non-RISC/6000 display,
because of its heavy dependence on the PostScript extension to the X
server. The Info Explorer, although very user friendly and powerful
when used on the local display, cannot seriously be considered as a
replacement for `man' for this reason. It is possible to convert much
of this information into standard man page nroff format, and I think it
would be a good project for the future. I don't know whether or not IBM
would be interested in doing this, but I'd suggest that it be mentioned
to them before much effort is put in on our side.
That's quite interesting... I bring it up on VAX/RT displays all the
time. It complains about a couple font substitutions, but overall, I
have not hit upon the sections that employ PostScript graphics. I guess
this will be a problem for some people and not others. In fact, I have
even displayed some sections that employ grids, so I don't believe they
utilize PostScript, in general.
---------------------------------------------------------------------------
To continue the discussion about "info", I should also add the following:
- It seems to be very anti-social with its use of the font-path, and if it
is not first, "info" looks terrible. If it is first, everything else
looks terrible.
- The hypertext database is 145MB, and most accesses to the man pages, and
other information have the tendency to flush the AFS cache of your
session. The default cache is about 40MB, but there seems to be no easy
way around this problem. So performance on the RISC/6000 is terrible as
you have to look up more information. If there were sufficient space
to replicate this database within NFS filesets and have those on the
individual subnets, performance may improve.
-R
