deadbug, I have yet to see a real naysayer. We're all just trying
to get a more concrete idea of what traked really wants to do. He
says he'd like to start a new distro, but the question still stands,
what will make the new distribution different from the many others
out there? The reason people think there are too many distros is
because there ARE too many poorly defined distribution projects
out there. If half of those working on the 96 distros instead spent
their time working on fixing what they saw as wrong with other
distibutions, then those, fewer, distros would be even stronger.

I'm not against free choice (If you knew me, it'd be like saying
Richard Stallman is for closed source commercial software),
I just think we can help make Linux better by improving what is
out there, rather than spending a few months of time working on
something that could end up just like 30 other projects -- with
a few good things that are hidden by the morass of poorly
defined goobly gook that's yet another distribution.

I'm not saying it's a bad idea to start a new distribution of Linux,
I'm saying it's a bad idea to do so without some very well defined
goals for the new distribution, goals that will make the distro so
much different from others that it'll be worth starting a completely
new project.

NSKL: the package tool for Slackware is, in fact, a very nice
project, the only problem with it is that there isn't a feature that
allows one to download the source from whereever it's needed,
compile it and install it. There have been a couple of attempts
at fixing that (autoslack, autopkg), but there still doesn't exist a
good database of packages and their dependencies, with
something like an rsync feature for downloading. There would
also need to be a way of checking for those dependencies on
the user's system. . .

That is a good idea

Moses: I was thinking about what you said, and creating such a database, or using an existing one (such as the debian one) is not an option i like. But i think a good idea would be to create a tool, something like installwatch, but specific for slack, that will create a slackware package from source. So people download a source of a program, tar zxvf'it, cd to it's base dir, and then the user can execute this new tool/script that will ask for any configure option you wish to put (and perhaps display ./configure --help to show the available options) then go on to compile and make the package, and make it using makepkg, and finally install it with installpkg, leaving a packagename.tgz in the source base directory. Later this package can be uploaded to linuxpackages.net. I dunno if this script would be useful at all, but as i learn more bash scripting i will attempt to create something like this, perhaps some people might find it useful, and if i get more people to use it, we will see more .tgz packages around, beacuse it will be easier to create them. Also, this way it will be easier to uninstall software from a slack installation (with removepkg). This will increase the use of pkgtool since all packages would be compiled from source, but installed as slackpacks, and will appear in pkgtool's package database on every system.

Does anybody have any other ideas on improving slack package management? I'm pretty happy compiling from source, but if make uninstall script is missing, it can be a pain to remove packages from slack. Also new slack users don't like source. So they can download pre-compiled binaries from linuxpacgaes.net. This always leaves the dependency issues, but this is easily solved with ldd and downloading other needed packages.

Now this is plain talk, and im so far behind with knowledge to create something like this, but im willing to try sooner or later.
ehrm... did all this make sense?
-NSKL

Moses, i completely agree with what your saying...
but the point is, I want a project team going so that we can think up ideas and make goals to make this distro unique.

I have put forward some ideas that I thought up at that moment in time which I think would be quite useful to the average Linux user.

I also understand that there is alot on my plate, but rather than rushing something out just to get it done, we can work it out and perfect it. Getting it done for a "set release date" is not really what I had in mind in the first place.

I toyed with the idea for a Slack package manager. I am actually still working on it. It will probably never get finished, and even if it does it will be too deviant from the Slack ethic for most people.

It is a multi tier p2p structure with different layers providing different services necessary to assemble the components and verify that they are valid. One tier to keep/submit md5 checksums and descriptions. Another layer to retrieve the files. And a third (optional) for dependency checking and compilation. It is actually fascinating from an intellectual standpoint, and even an achievable goal since 8.1 (with the inclusion of version numbers on the packages).

The hard part is dependency and GCC version checking. Since by it's nature Slackware does not, and never intends to, include dependency checking it would need to be added on externally. Then comes the question of how do you establish dependency? based on the libraries, or upon the package? And to what extent do you need to concern yourself with versions and under what version of GCC the library was compiled.... I think you see my conundrum. The p2p network is easy, the semi-securable MD5 network (based on signatures from known peers) is achievable, dependency checking on the other hand is a little more tricky, for me a Nth difficulty problem.

P.S. Sorry for being off topic, but I had to mention my thoughts on the subject...

Hey,

I am not much of a coder, so I can't really help you there, but I would make a great alpha/beta tester. I am quite good at web design (isn't everyone!) including css/php/apache/mysql and all that jazz so I could help with the web side of things. I am also an accomplished writer, so I could help do the documentation...

Email me if you want more info.

This is the sort of project I would love to get in on...

Oh yeah, a more automated way of creating the tgz packages would be nice as well. A chroot jail is the easiest way I have found but it is not necessarily straight forward.

Mephisto, NKSL, I've started another thread to discuss the
slackpack manager. . .

http://www.linuxquestions.org/questi...threadid=40410

I'm still interested in the new distro. I really don't want to discourage
anyone from spending time on making linux better. . .

I also had the idea to start my own distro, since my first learning computer for linux was a toshiba 1910 laptop with a 200 MB HD and not cd rom, only floppy. i wanted one with a minumalist feel, only text, and maybe xfree86 1.0 if at all.

unfortunately, im already working on another project for blade runner, WWW.bladezone.com, any fans interested ?

i CAN however, aside from working on my blade runner project, i can help out in small ways in this distro if you'd like.
DAChristen29@aol.com

Just a note for those wanting to participate in the development of this (or any other) new distro. You do not need to be a programmer to make a distribution. Someone on the team will need to be one to resolve issues sure but most of the work involved in creating a working linux distro is not coding. IMHO the most valuable person on a team of this sort will be a systems administrator type. Someone who is good at compiling packages, not writing them. Someone who is good at keeping track of the myriad of details involving all the different packages. Someone who can document the structure and status of the entity.

So if you are not a programmer but are good at detail oriented tasks such as those a sys admin does on a daily basis, believe me you can be a very meaningful part of a team.

(Sorry about the previous OT posts.)

traked,
You could think about starting a gaming distro. No distribution so
far has what it takes to really grab game developers' attention.
Mandrake was making noises about it, but they're in trouble, and I
never saw any notable results. . .

Hey

I want to start my own distro to, trying to get LFS going here, having a little problem with it though, but I hope I can get it done. Now I am fairly new to Linux, only about a year and a half, but I am willing to learn more about Linux, and I think the best way would be to work on a distro.

Now that is an excelent idea.

A gaming distro is something that i have not seen done anywhere to any proficency, to see a gaming distro would be great, i cant code and am slowly learning more and more about linux as the days fly past, please mail me if you want some help and consider going down the path of a game distro

Any type of distro that address a specific area could be a contender. Most of the big distros try so hard to be something for everyone the pure size and complexity is approching critical limits. 7 cd's with Suse 8.1 pro? Get real how many average joes will ever use half of that.
Decide on a niche gaming or whatever and just use what you have to. Don't use all the 75 different programs to do the same thing so people will have a choice. Make it lean and mean.