1 - It has a "wow" effect for first users and it is easy to remove the unwanted application. There has been a lot of request for a minimal pdaxrom rom. It should not be too difficult to build such an image with the builder, one question is how small should be this image. only console with libc and bash? minimal X?
As name is pda
Xrom, I think image should consist of X11 + system utilities and configuration tools.
2 - I agree, I think the main problem here is time, you need to test the upgrade process, track the changes, write more intelligent scripts in the ipks that take into account previous ipks etc ... It is much easier to reinstall from scratch.
Reinstalling is not easy process. Much easier "ipkg update; ipkg upgrade". I know, it some harder for developers. And this is because of used "builder" which build all used software at one time.
When main efforts will be spent to core (kernel, X11, libc, locale...), userspace software can be maintained by other people. Indeed, I think, if sashz & laze will be free of support of gnumeric/abiword it will more usefull for process.
All they will need is to support kernel/libc.
3 - there is unclutter, as a workaround here www.pdaxrom.org/unstable. Though I agree it's not very usefull, and even cause me some trouble when
the cursor is on the way, I like to know where it is. I should try to use unclutter to see If I really miss the cursor or not.
Why do u need cursor at all? I don't need to track cursor, because I alwas can click on needed place (there is no cursor in any other systems -- palmos, wince, cacko, u-name-it).
4 - Interesting, have you tried to use it? does this change the size of the executables? Does it really make a difference?
I never tried it on ARM (can't get compiled). But use it on dekstop. It change size (exec some bigger). Result executables are faster to load.
