Re: [Mentalunix-developers] Read

[Rob B. <ro...@fu...>]

On Thu, Mar 15, 2001 at 04:34:38PM -0500, Clinton Ebadi wrote:
> I've really been thinking about this for a long time now. After 
> looking at the list of Debian project goals, and our project goals, I 
> found something odd. 
> 
> Here is a list of our goals, with a '-' for debian not having the 
> same goal, a 1-9 for how much it is the same as ours (10%, 20%, 30%, 
> ....) and a + for being the same as our goal: (the rating is at the 
> end of the line)

Just to update you here on Debian

> 
> Easy installer	9

Best installer i've ever used but its currently being rewritten.

> - autodetects hardware +

Pregeny is giving this feautre back - easy to implement using discover

> - Graphical mode +

Progeny - GNOME Debconf frontend most of the work

> - text mode +

Hmm ala dialog?

> - Configuration for cd burners and other misc devices -

So much work to be done, but i agree could be made easier
 
> Standards Compliancy +
> - FHS +
> - LSB ? (LSB Sucks anyway)
> 
> Package Managment 7
> - Support for all formats 2 (through alien, a messy way)

Debian is moving to udeb which is a new format for its new installer

> - Rebuild a full tree of source for any target with 1 command -

Hmm, simple ickle script! new suggested CVS will also help this

> - Automatic Dependcy tracking + (apt has been around for a while now)

> - Graphical Package Browser +

Quite a few under development, Netsnipe showed me one the other day that was
great!

> - Easy tools to make packages -

You blind? debmaker and debhelper jsut cd to a autoconf/automake software
distribution and type debmake or debhelper
 
> User Interface 9
> - Complete Desktop Enviroment 9 (a few small differences, non major)

Huh? GNOME/KDE/WMaker take your pick

> - Easy GUI Config +

Pretyy easy

> System Administration 8
> Self maintaining + (once its up, it stays up)
> Tools to ease admin 5

Done your research? Debian has the largest number of great admin tools i've
ever seen. Deborphan for instance is coool

> overall 9
> 
> If you look at the things we differ on, they are stuff that can be 
> implemented at the application level. Take Tools to do system admin --
>  we could help with the push to debianize linuxconf and possibly 
> write our own system admin tool. For the unified package manager, we 
> could make mpkg a frontend to alien + dpkg (if it sees that the 
> package is rpm, it will alien it, then use dpkg). I think it would be 
> cool if the debian developers accepted stuff like mpkg to "replace" 
> dpkg (meaning, you still had dpkg but stuff like apt called mpkg 
> instead to allow for rpms to be installed). 

Thats a lame idea, why would debian users want to use crappy rpm packages
when deb and udeb are much better. Why recreate work?

> So, why do an entirely new distro, when we could focus on the 
> application part of mentalinux, making a great distribution even 
> better. It would probably get us more publicity and have more people 
> willing to help out. So, now I lay out a plan that you can choose to 
> accept or reject. This is a democracy -- if a majority wants to do a 
> complete distro, then we will.
> 
> 
> First, we need to find a server that is running debian, and would be 
> willing to allow us to use its servers. We would need httpd,  
> annonymous ftp, and some dpkg / apt stuff to set up the package 
> mirror. 
> 
> All the packages we would package would be built against debian 
> unstable (am I the only run running it? My box is really slow, so the 
> builds might lag a few days from the source if it just me). A task-
> mentalinux-core and task-mentalinux would be set up to provide all of 
> our packages and just the core ones (mpkg and such). We could just 
> start off by popping some early dev tarballs of each app until I get 
> the hang of making debs (help me, trusty debian packaging guide). 

Heh, thats horrible to read!

> In addition to just the core apps, we would also provide some 
> packages regular debian doesn't. As you might know, it is a bit 
> difficult to get into debian (and more so now). So, we could let 
> people on to the mentalinux project with less work (just ask, show us 
> a package, talk on IRC for a bit, basic net backup check, etc). 
> Mentalinux won't have to be a through as debian is, since we will 
> have less developers and just focus on a few packages. That should 
> make tracking down bad packages a lot easier. 
> 
> After debian woody its released, we can try to pacakge some isos with 
> the mentalinux addon stuff, and mess with the post install part of 
> the installer to allow the mentalinux config stuff to take place. 
> Example: the user would install all the packages they need, then quit 
> dselect, wait for the stuff to install, then the mentalinux config 
> app would come up. It would allow the user to configure stuff like cd-
> r drives and other devices the actual installer doesn't take care of. 
> 
> 
> After we have proven that mentalinux has good things to offer, we can 
> think about merging with debian. In the meatime, the ISOs and stuff 
> that we would distrubute (after a new stable is released) would be 
> Mentalinux-enhanced Debian GNU/Linux. I'm guessing that the woody 
> freeze process is going to take awhile, which is why I'm aiming for 
> us to release a stable release (althought maybe not 1.0 of 
> everything) when it is ready for release. 
> 
> I know this plan is a bit incomplete. But, I'd like you to consider 
> it. So, to make our own distro or to work on debian? It's all up to 
> you (my vote is for, of course, enhancing debian)

My view is just join Debian as maintainers and help develop it that way,
after all why clone the work?

rob 
-- 
Rob 'robster' Bradford
ro...@de...

Founder/Lead Developer: Debian Planet