1: [[!template id=project
3: title="Move beyond TWM"
13: Right now if you install NetBSD with native X11 you get the same
14: default X session that X has shipped forever, complete with TWM. Just
15: like it's 1989.
17: Obviously this behavior can be improved after installation by adding
18: suitable packages and choosing a different window manager, or
19: installing one of the full desktop environments. Veteran NetBSD users
20: know how to do this. Veteran Unix and X users who aren't veteran
21: NetBSD users can figure it out pretty easily. Most such people have an
22: X setup they've been using for years that they can and do just plop
23: into place on a new machine.
25: This project does not affect those users, even the ones who for some
26: crazy reason *want* to use twm. It also does not affect users who just
27: want to install GNOME, KDE, or whatever other desktop.
29: The goal of this project is to provide a less knobby default X
30: environment for users who are willing to run plain X rather than a
31: full desktop but do not have their own X setup.
33: The project is: choose a window manager suitable for replacing twm in
34: base; import it into base; and update the default X config to provide
35: a decent default session using it.
37: Note that it needs to be a mainstream window manager; tiling window
38: managers and other exotica are not really suitable for the purpose. It
39: should also not depend on anything that isn't in base (that is, base
40: with native X), it should be smallish and not a bloatmonster, and
41: ideally it should be BSD-licensed.
43: Based on the window managers in pkgsrc, there are two chief
44: candidates: fluxbox and golem. Each has pluses and minuses. Some of
45: these are:
47: * golem is prettier and fully themeable
48: * golem is written in C, fluxbox in C++
49: * golem is a lot smaller (less than 1/10th the code size)
50: * golem currently needs some work, chiefly on the config system; fluxbox appears to be ready to go
51: * golem is just about dead upstream; fluxbox is not very active but is still getting occasional releases
53: Currently it looks like golem is a better choice, especially since
54: ideally we'd like to pick something that will run decently on
55: older/slower ports. But it requires more work.
57: Note that we could (AFAIK) become upstream for golem; this is not
58: necessarily a bad thing.
60: There might be other possible choices; these two are the ones in
61: pkgsrc that seem viable.
63: Note: this project is not listed as a GSoC project (although doing the
64: needed work on golem maybe could be one) because much of it has to do
65: with build system integration and ancient X things and other stuff
66: that requires experience a student probably won't have.
CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb