[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: custom vs. derivative (Re: packages.gz corrupt, missing packages and other issues)

On Mon, 31 Mar 2008, vagrant@freegeek.org wrote:

while i fully support pushing to include everything entirely within
Debian, i think it's important to acknowledge that *sometimes* it may
actually be better to develop a proof of concept outside of Debian

Yep, no problem about this.  That's why I wrote that we are tolerant
to use the list for this.  My only problem is that I never met
anybody who perceived the meaning we once gave CDD and I'm so bored
to explain over and over.  I think even XYZ (without having no
meaning behind these letters at all) would be a better name.

i think debian-edu (a.k.a. skolelinux) is a fine example of this, and
it's clearly made huge progress towards full integration within Debian
proper, but it still doesn't meet a strict definition of a custom debian
distribution (a handful of packages and the CD/DVD images are outside of
Debian still). but the intention is clearly and genuinely to integrate
into debian, and that's what i think really matters.

Fully ACK.  And BTW, SkoleLinux was before there were Custom Debian
Distributions and thus they did not even had a chance to become one
from the beginning.  Chances today are better and sometimes people
just do not know this chance.  It is our fault if we would not tell

working outside of debian has has pros and cons, the biggest con being
the very real risk of good ideas never actually making it back into
Debian. that's not to be taken lightly. there's also the possibility of
duplicated work.

Also fully ACK.

on the other hand, i think you can't also disregard the power in the
enthusiasm of getting something working *right now* vs. waiting for some
changes to get accepted within Debian, which can be an unreasonably long
wait sometimes.

Yes, that's why I try to support those people here if I can, but
I try to express my opinion based on past experience as a hint for
newcomers.  That's all.  I have neither any means nor power to
force anybody to do something else than he wants to do. ;-)

waiting for bug fixes, some new version or new package
uploaded, slow buildd's stalling progress to testing, whatever ... can
discourage people from making good progress that might be integrated
into Debian.

I perfectly know the reasons for this approach.  There are even
quite successful distributions born out of this.

either way, we *might* loose improvements to Debian.

Sure, but we are losing many things without notice.  But if I see
a slight sign that we might lose something I try to take the chance
to at least point this out.

there are so many similarities in the issues that derivatives and custom
debian distributions face that it seems important to me to encourage
cooperation and collaboration between the various projects, even if they
do not make the same choices and priorities.

Reminds me about Agnula.  It was a shiny project sponsored by EU.
What happened to it.  I don't know ...

in the end, yes, it's *great* to fix it within Debian, but i hope we can
envision ways to help derivatives transition into fully fledged custom

Yep.  My way to help is making CDD so attractive that they want to.
I'll show you the web stuff once we meet the day after tomorrow. ;-)

BOF at debconf8, anyone? :)

https://penta.debconf.org/penta/submission/dc8/event/244 and

live well,

See you in Extremadura



Reply to: