Re: CDDs projects on Alioth
|--==> Andreas Tille writes:
AT> On Sat, 9 Oct 2004, Ben Armstrong wrote:
>>I strongly disagree. All CDD projects will *never* be in Alioth. You
>>cannot force everyone to use Alioth.
AT> I do not want to force anybody who is maintaining free software and I do
AT> not want to. But I obviousely did a wrong statement in the docs (which
AT> more or less says that the CDDs can be found in the cdd tree on alioth).
AT> This has just to be fixed - no more no less.
Well, when I wrote about those 30 derivative distros, I did mean that
I expect all of them to put their sources on Alioth or integrate with
Debian. It was just an example to show that keeping all present and
future CDDs in the same project can eventually be problematic, given
the number.
I absolutely agree that everyone should be free to place his CDD
wherever he wants.
I was just wandering if we should have some guidelines for at least
*our* projects, as CDD mailing list members and Debian insiders. Of
course anybody is free to stick to them or not.
>>If even one project does not put
>>their project there, you won't be able to say in the doc "all CDDs are
>>here" pointing at an Alioth category page. Such an index should simply
>>be a manually maintained page listing all known CDDs.
AT> My idea was that having CDD on a common place would make them easier to
AT> view. If people think that this is not the case - it would not be the
AT> first time that I was wrong. However, if I had talks about CDD and I
AT> will be asked by the auditorium I have a very strong feeling that pointing
AT> to a single place makes a well organised impression and thus I keep on
AT> thinking that having one place has advantages . Nobody was able to
AT> convince me to find a better approach yet. But I do not force anybody -
AT> I'm just stating my point of view and recommend what my experience seemed
AT> to make a logical decision according to observing our users expectations.
Yes, having a common place *does* help, but I think this can be
accomplished by having different Alioth projects rooted in the same
directory. This assures both visibility and independence.
However I agree that using the current Alioth CDD project as an easy
start up point for new CDDs might be handful. When and if a new CDD
grows enough to need a separate project, than the team may consider to
open its own Alioth project. Again, in such case I think it would make
sense to have all of them under the same directory.
>>Please don't confuse children (people) with children (descendants in a
>>tree of derivatives). The above URL refers to the latter. Maybe you
>>haven't and you really mean the latter, but mentioning children in both
>>senses almost in the same breath blurs the distinction.
AT> Uhm - thanks for the clarification. I was sitting behind a slow connection
AT> and did not visited the link which would have changed my arguing. ;-)
>>But yes, as Free has said, keeping all CDDs under the Alioth CDD project
>>is not sustainable.
AT> As I said: If there ar reasons for a different organisation it is fine
AT> for me. We just need a kind of organisation that ensures that we can
AT> find a single sentence which, spoken to a newbe, explains, how he can
AT> find all CDDs without any effort. If this is possible (and this sentence
AT> moved to the docs) I'm completely happy.
>>And yes, as you say, Andreas, there are many
>>distros with only subtle differences that could probably merge. But
>>don't overestimate how quickly that will happen.
AT> Well - this was my misinterpretation of "children-distros" ...
>>Look at the
>>proliferation of packages with similar purposes in Debian. I don't see
>>a whole lot of them merging, not the least reason for which is that
>>every group of developers has their own ideas, goals, and comfort
>>zones.
AT> Sure - I do not want to stop them: I was just wondering about 30 different
AT> distros only for children as target users ...
>>Again, Free, it is fine to have a recommended approach (creating your
>>own Alioth project) and an alternate approach (making it a subtree of
>>the CDD project). Do you really think all 30 of the above, even
>>assuming they wanted to switch to Alioth, would opt for the latter? I
>>don't. In fact, I think it will be rather rare. If and when it starts
>>to become a problem, we can deal with it then.
AT> I really think that most of them are not really fullfilling the
AT> category "portantial CDD". Half of them are commercial and thus can
AT> not be integrated into Debian and many others have reasons to stay
AT> outside. I see potential new CDDs coming from inside of Debian in
AT> terms of specializing for special fields or locales.
IMO the great challenge for Debian is how to let even commercial
derived distribution to be as much as possible regular CDD. That would
be great source of man power and quality assurance, and everybody
would benefit it. My understanding is that this what the UserLinux [0]
project is all about.
Cheers,
Free
[0] http://userlinux.com/white_paper.html
Reply to: