[2529] in Release_7.7_team
Minutes of 2000-01-03 release team meeting
daemon@ATHENA.MIT.EDU (Greg Hudson)
Wed Jan 3 16:56:48 2001
Date: Wed, 3 Jan 2001 16:56:44 -0500
Message-Id: <200101032156.QAA30871@equal-rites.mit.edu>
From: Greg Hudson <ghudson@MIT.EDU>
To: release-team@mit.edu
Attending: ajfox ghudson rbasch edolan amb wdc jweiss
1. Matlab issues
Matlab 6 comes with a big Java-based IDE. A couple of problems with
it:
* On Linux, the print command seems to look for /usr/bin/lp.
We may need to make a symlink.
* On Solaris, the IDE seems generally buggy. They suggest we
should get patches for the JVM from Sun.
The third-party group should look into the print command and see if
the behavior can be changed reasonably before we add another hack to
the release.
Greg will look for Sun patches for the JVM, since miki is out of town.
2. AUI config
(A) Sawfish themes
Greg presented a handout showing six candidate themes (microGUI,
brushed-metal, TigDomeII, gtk, SawBoxL, and helix-gray), the latter
four of which were mostly selected for low color usage. Comments
received:
* It's probably okay to have a color-hog theme as long as it's
easy to change for people who still use 8-bit displays.
(This means there might be better choices than the six
themes Greg chose.)
* microGUI is pretty, but the nonstandard minimize/maximize
buttons are a big lose.
* helix-gray is fairly nice and low on color usage, but its
buttons don't actually look like buttons, which was a big
concern for Bill and a smaller concern for others.
* Bill prefers brushed-metal, but a number of people find it
rather ugly.
* SawBoxL is king of nice except that it's missing sides.
What we should probably do is narrow down the choice to ~3 themes and
see if the usability team can find out what users would prefer.
[Though in retrospect, I'm not sure if aesthetic issues are really
fodder for usability testing.] Of the six presented themes,
brushed-metal, SawBoxL (or something like it), and helix-gray seemed
like the three best candidates.
A mostly orthogonal issue is whether to move the close button over to
the left, replacing the menu button, to get it away from the minimize
and maximize buttons. Choosing to do this would limit our theme
selection to themes which use the same color patter on the left and on
the right, or to themes which use semitransparent PNGs, or to themes
which already happen to have the button there; otherwise we can't move
the buttons without problems.
(B) Panel
We pretty much can't modify the panel properties once we decide on
them; we could only reach new accounts. So we're stuck with the
launchers and applets once we decide what they are.
Consensus was that we should have an editor launcher which should (for
now) run emacs. It's questionable whether we will ever want a word
processor launcher.
There was some discussion of Greg's menu superstructure plan (global
main menu with programs and desktop properties on main menu, favorites
as a submenu), without a complete consensus, particular as to
user-customizability of the menus. There was consensus that any plan
which does not allow us to add and remove menus from the top-level
dash-menus analog is a non-starter.
(C) control-center
The only issue here is whether to support capplets like
background-capplet which would have to be initialized at login. The
plan is to support them unless they make the login sequence too slow
on the slower Suns.
(D) gnome-terminal
We would like to use it in preference to xterm. The concerns are:
* No option to reset the terminal. (Maybe we can hack it in.)
* Changing the font size is harder since it doesn't have a
nice menu of fixed-width fonts like xterm does.
Assuming we do use it, the default font size has to change. We should
pick the largest font such that it is possible to have two
side-by-side 80 column windows on our narrowest cluster resolution as
of this fall. We're not sure if the narrowest resolution is 1152x900
or 1280x1024.
3. AUI testing plan
It's too hard to make the Athena source tree fit nicely into a locker;
different assumptions are made. (prefix vs. exec-prefix, which I
guess we could work around by not sharing the sharable parts, but also
various assumptions about /usr/athena in hacks to third-party
software.) So the testing plan should revolve around bleeding-edge
Linux-athena.
4. Other AUI issues
On IRIX, we were planning on sticking with 4dwm unless the user sets
skip_sgi. But we were still going to replace dash with panle. This
could lead to some collisions between the 4dwm universe the gnome
universe, e.g. who gets to set the background and what happens if the
user tries to configure it the wrong way.
Bill reports that Jeff thinks we absolutely positively need a menu
option to back out to the old interface. The release team has some
concerns about what that says about how long we'll commit to
supporting dash and mwm. Probably the best place to address this
would be an integration review, which we can do when we have things
mostly configured in the source tree.
5. OS upgrades for 8.5
amb should have time to do the Linux upgrade.
6. 8.4.18 plans
We will aim for the dev cell tonight and the Athena cell on Jan 29.
