[lug] The original Debian announcement

[Nate Duehr]

Saw a note that Debian is 15 years old as of 16 August 1993, and was 
thinking about the goals of the project when it started:

http://groups.google.com/group/comp.os.linux.development/msg/a32d4e2ef3bcdcc6

Please note, I'm a Debian FAN.  I'm not slamming it as part of some 
childish distro-war mentality.

I'm just looking over the project's founding goals versus where it's at 
today, and commenting on what I see.

Ian Murdock's original announcement with comments:

----------

What will make this release better than SLS?  This:

1) Debian will be sleeker and slimmer.  No more multiple binaries and 
manpages.

*** That didn't work out so well, did it?  20,000 packages and climbing.

2) Debian will contain the most up-to-date of everything.  The system 
will be easy to keep up-to-date with a 'upgrading' script in the base 
system which will allow complete integration of upgrade packages.

*** Up-to-date, nope.  In fact, Debian has become one of the slowest 
distros to upgrade packages, and has culturalized the process to say it 
creates a better quality end-result.  Interesting that is NOT what the 
original founder wanted, I think.  Easy to upgrade, yes... they got that 
one nailed compared to most distros when packages don't make major 
architecture changes.  (Which happens, so often that ... maybe "ease of 
upgrades" is a failure after all, but not really Debian's fault.)

3) Debian will contain a installation procedure that doesn't need to be 
babysat; simply install the basedisk, copy the distribution disks to the 
harddrive, answer some question about what packages you want or don't 
want installed, and let the machine install the release while you do 
more interesting things.

*** Hmm.  Yes or no?  The number of questions has grown so much that it 
sure FEELS like I'm still babysitting the installer.  The topic of how 
to FULLY automate Debian installations comes up often on their "-user" 
mailing list and the best ways seem to be a separate project called 
"FAI" or the various imaging tools available today.  So, MAYBE Debian 
met it's original goal but... It doesn't feel like it.

4) Debian will contain a system setup procedure that will attempt to 
setup and configure everything from fstab to Xconfig.

*** Kinda.  The setup can't keep up with the complexity of the large 
packages, though -- try REALLY configuring Apache or X from only the 
Debian setup tools... no way.  dpkg-reconfigure... helps MOST packages 
gain a bit of initial setup sanity, but after that you're a babe in the 
woods.

5) Debian will contain a menu system that WORKS... menu-driven package 
installation and upgrading utility, menu-driven system setup, 
menu-driven help system, and menu-driven system administration.

*** Hmm, package installation was dselect back then, and aptitude now, 
and dpkg under the hood now... so okay we'll give 'em that one. 
Menu-driven system setup?  Not after install.  See above.  Menu-driven 
system admin?  Not a chance on a system doing anything really 
important/useful with large packages/systems.  There's no "menu" to see 
  how your Tomcat server is doing, and never will be.  Not an impossible 
goal, but not likely to ever work correctly either.  (Someone could 
mention something like Webmin here and I'll barf... plus Webmin isn't a 
Debian-specific project.)

6) Debian will make Linux easier for users who don't have access to the 
Internet.  Currently, users are stuck with whatever comes with SLS. 
Non-Internet users will have the option of receiving periodic upgrade 
packages to apply to their system.  They will also have the option of 
selecting from a huge library of additional packages that will not be 
included in the base system.  This library will contain packages like 
the S3 X-server, nethack and Seyon; basically packages that you and I 
can ftp but non-netters cannot access.

*** Hmm, considering the history -- okay yes, packages were offered of 
things that were ONLY available online back then.  But if the goal was 
to CONTINUE to offer a system that could be kept reasonably updated via 
off-line packages... that one has to get a failing grade.  The packages 
are far too big, and updating from a CD or similar these days would be 
an enormous pain in the ***.  Both because it's not a priority anymore, 
and also because security packages come out as fast as I change 
underwear these days, thanks to crappy coding practices that haven't 
been somehow dealt with in 15 years, I don't think this one matters too 
much anymore.  But, honestly it doesn't matter because if you're not 
connected to a network, most security packages are not needed anyway... 
unless you've got users that need to be smacked around a bit (local root 
exploits), and you probably know where they live, anyway.  Drive over 
and beat them with a stick.

7) Debian will be extensively documented (more than just a few READMEs).

*** Uhh, maybe a passing grade here.  Lots of documentation, little 
focus on whether any of it is useful or up-to-date.  The web has turned 
into the "documentation of choice" for almost everything, but I will 
admit that Debian's /usr/share/doc directory usually contains a lot more 
useful information than some other distros.  I wouldn't call it a 
shining example of the best software documentation I'd ever seen, but 
it's "passable", maybe.  Depends on the package and the maintainer.  No 
consistency.

8) As I put together Debian, I am keeping a meticulous record of where I 
got everything.  This will allow the end-user to not only know where to 
get the source, but whether or not the most recent version is a part of 
Debian.  This record will help to keep the Debian release as up-to-date 
as possible.

*** Seems like a revamp of #2 in many ways, but it's kinda interesting 
that 15 years ago, "distros" didn't document where they got things from 
upstream.  Amazing considering the mess we're all in re: Licenses 
nowadays, isn't it?


9) Lots more, but I'll detail later...

*** A common theme in life.  Translation:  "I just ran out of time 
and/or energy."

:-)

---------


So... what do you guys/gals think?  The original goals of the project?

A+?  C-?  F?  What's the grade?

You guys have seen my comments about "quality vs. quantity" lately in 
regards to Linux... did Debian stray away from Ian's original intent?

Further comment; thoughts -- the current "Social Contact" is here:
http://www.debian.org/social_contract

Also I found it telling and/or sad, that not a single named person is 
listed as being a part of the Quality Assurance team within the 
organization:

http://www.debian.org/intro/organization

Interesting stuff... looking back 15 years... isn't it?

Nate