[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [school-discuss] Linux distribution



I'm glad I got some discussion going on the list.  If anyone would
like to discuss distribution design or what distributions are lacking
further, feel free to join in at the Eerie Linux forum.  One of the
reasons I mentioned the project is that I'd really like to re-examine
some of the assumptions that most Linux distributions make and would
like to brainstorm design ideas with others.

I think I've mentioned it before, but I have my own ideas of what I'd
want in a Linux distribution.  I've tried several distributions.  I've
been involved with some projects and have tried to get involved with
others.  So far, nothing's worked out.  Don't know if what I'm looking
for is so far off the map from what everyone else needs that there's
no interest or if I just haven't hit it lucky yet.  My main goal is
for a lightweight distribution that could be built from scratch and
understood and debugged by one person.  I'm concentrating on
lightweight applications that work well on older systems and are
efficient on newer machines too.  I feel like many Linux distributions
lack consistency.  Some are better than others though.  Some tend to
install files wherever the application puts them.  For example, some
honor the HOME directory for setup files while others have switched to
XDG_CONFIG_HOME.  I want applications to work together and be
consistent.  I have been looking into educational applications, but
mainly applications using lighter GUI libraries.  So, programs like
Tuxpaint are definitely of interest, but larger more bloated programs
aren't even on my radar at this point.  I'm also carefully looking at
licensing and patents.  I want the results to be something there's no
problem distributing.  I've been looking at Red Hat's version of xine
which is even more limited than Debian's version.  Documentation will
be a very important part of the system.  I plan on integrating it with
the build and package management system.  My goal is to include how to
type documentation with examples of how to do tasks I commonly perform
on my system such as how to create a CD or DVD with my own content.
(I should mention, there are a few Linux systems that come to mind
that try to have user friendly documentation for new users like
Absolute Linux.)  I'm also looking at common libraries and programs
and investigating whether there are alternate implementations that may
be more efficient or better designed from a coding standpoint.  I'm
using pkgconf instead of pkg-config and have replaced packages like
GNU gettext with other compatible alternatives.  Those are just my
personal goals and I am working on implementing them.  However, I've
been working on this for quite some time now and I'm nowhere near what
I'd considered a release phase.  I don't believe in release early,
release often.  I like my programs and libraries well tested and
stable before I release them.

I'm finding it very interesting that the typical response when I
contact a project is if you want it done, go do it yourself.  Now,
that I'm approaching projects and Linux users with the idea that I'm
doing it myself and rewriting a lot of code I'm not happy with, the
typically response is 'Why are you doing that?'  When I volunteer to
write or rewrite the code myself, they act like I shouldn't do that.
You can't have it both ways.  If a project doesn't quite fit someone's
needs, someone's going to rewrite it.  That's one reason there are so
many forks of Open Source projects out there and with source control
systems like github which are designed for this sort of thing, there
will be even more.

One of the other reasons I was interested in the Eerie project was the
possibility of sharing bits and pieces of what I'm working on with
others at this point in time rather than waiting to share a more
complete project.  For those like me who have yet to find an Open
Source distribution they're comfortable with, I can certainly
sympathize with that sentiment.  I see people distribution hopping and
trying new distributions all the time and if users had everything they
needed in their own distribution of choice, I can't understand why
they'd need to do that.  If there are design goals you'd like to see
in a distribution, would love to discuss this further either on the
forum or even by e-mail.  (I have a C/C++ design mailing list for
discussing more technical program design issues too.)

Sincerely,
Laura
http://www.distasis.com/cpp
###
To unsubscribe from the schoolforge-discuss mailing list:
Send an e-mail message to majordomo@xxxxxxxxxxxxxxx with no subject
and a body of "unsubscribe schoolforge-discuss"