Re: [ad-hominem construct deleted]
On Tue, Jan 17, 2006 at 06:46:26PM +0100, Gerfried Fuchs wrote:
> * Matt Zimmerman <mdz@ubuntu.com> [2006-01-16 15:39]:
> > Is the meaning of this statement truly unclear to you, or is this purely a
> > rhetorical point? Under the assumption that you read it differently than I
> > do, I'll attempt to explain.
>
> Do we call RMS a Debian developer? Do we call Linus a Debian Developer?
> Does anyone seriously consider that?
Kennedy wasn't a citizen of Berlin, either, not literally. The world
understood what he meant, though, when he said (somewhat awkwardly) that he
was.
> Pardon, but that's ridiculous. I don't have upload permission at all,
> can't do anything about my packages, there are changed packages with
> still my name as maintainer that I never got any information about --
> and you still have the guts to call me a Ubuntu developer? Sorry for
> laughing into your face for that...
It isn't productive to take this kind of jeering tone.
> > Given that you saw this on a wiki page, a disclaimer about wiki contents
> > should be implicit.
>
> It's still as cite from Mark on there, and I don't think that the cite
> is wrong. Or do you rather consider your fellow developers putting false
> statements intenionally there?
I'm saying that you should pause and consider that you're looking at a
world-writable resource before treating its contents as a position statement
on behalf of the project, and that malicious intent is far from the only (or
even the most common) reason for errors. It could very well be that Mark or
someone else originally wrote "from Debian" and the quote was transcribed
incorrectly.
In any case, as I said, I think the meaning of the sentence as a whole is
sufficiently unambiguous, though for the sake of clarity I will ask Mark to
look and correct it if appropriate.
> > It was inappropriate for this user to raise this issue with you,
> > rather than with Ubuntu, but that's been discussed elsewhere in this
> > thread already.
>
> So? There is the Maintainer field that still has my name and my email
> address in it as being responsible for that very package -- where I
> can't do anything against it. That's simply wrong.
This had been commonplace for Debian derivatives for years before Ubuntu
existed, and when the issue was raised regarding Ubuntu, I asked for input
from the Debian community as to what to do. The issue is not at all
obvious, and in fact it's quite similar to the attribution of upstream
authors of packages which are modified in Debian, which is even older.
> > What I find interesting about your statement is that you seem to imply
> > that the situation would have been better if you had been notified
> > that your package was a part of Ubuntu.
> [...]
> Yes, the situation would had been _immensly_ been better. It would had
> shown at least that Ubuntu cares for its upstream.
Ubuntu has been in communication with Debian, primarily on this and other
Debian mailing lists, about what we are doing since before the project even
had a name. We've been very vocal about our development process, which
essentially amounts to a branch of the Debian archive. I don't think that a
credible claim could be made that Debian was not notified that Ubuntu
includes packages from Debian. This is what it means to be a derivative.
> > This would be technically simple to implement, but I'm not convinced
> > that it's possible to do it in a socially acceptable way. Emailing
> > every Debian maintainer to notify them that their package is present
> > in Ubuntu sounds like spam to me, and posting Ubuntu-related
> > announcements to Debian mailing lists has been deemed inappropriate by
> > many in Debian as well.
>
> From first I knew only that there is this Ubuntu which goes for one CD
> with gnome and xorg on it. I thought fine, I don't have a package in
> that range, so why should it bother me too much, so I didn't check. Do
> you really think that everyone in Debian is aware that there exist a
> thing like multiverse or whatever which seems to include every single
> package that is in Debian? I wasn't, for a very long time.
Debian, too, distributes software via networked mirrors which is not
included on the official CDs. There is nothing surprising or devious in
this.
> An announce along that lines instead of a press release so you can add
> d-d-a to your announce lists would hadn't stirred up so much bad blood
I haven't a clue what you're talking about here. What press release, and
how does d-d-a enter into it?
> > The creation of Ubuntu was *very* widely publicized, as was the fact
> > that it was based on Debian, and this fact has been mentioned
> > countless times since, both in the press and on Debian mailing lists.
>
> But it wasn't really mentioned that it includes every single package
> that is out there....
Ubuntu is, and always has been, a branch of sid. This has been pointed out,
among other places, on debian-devel and on the front page of LWN. Not a
subset or a miniature distribution, but a derivative of the complete Debian
system. If you had doubts about which packages were included, it wouldn't
have taken much effort on your part to find out.
> I ask you, Matt, because you seemed like a reasonable person at the
> debconf: Do you think I'm absolutely wrong with that? Putting out
> changed things in the name of others, not informing them, with them not
> being able (besides not knowing about it) to change it?
Do you truly see this as such a radical departure from how Debian and other
distributions already work? Free software is rarely so clear-cut. By the
time a piece of free software arrives in the hands of a user, it has passed
through more than one set of hands and more often than not, modified from
its original version. This is not the result of an evil Ubuntu conspiracy,
but common practice in the community, and I do think that your reaction is
somewhat exaggerated.
As soon as the issue was raised (and although it was raised in a Debian
forum, without any attempt to contact a representative of Ubuntu), we
responded and engaged in the discussion. In the end, the discussion seemed
to die out without a clear consensus, and so it was dormant again until now,
where we're seeing some of the same arguments without regard for the
discussion which came before.
> > Setting aside the debate about patch submission for a moment, in the
> > case of most packages, there are no patches in Ubuntu relative to
> > Debian.
>
> So you even confirm with this message that the pulling job that seems
> to be expected from Debian Developers is a stupid thing? At least here
> we agree.
I don't agree there. It may not be ideal, but it certainly isn't stupid,
and the debate isn't about pulling vs. pushing anyway (it's easy for a
client to convert a pull mechanism into a push mechanism), but rather about
how much personal attention can be reasonably applied to patch submission.
> > - Packaging transition for the gcc4 C++ ABI. Debian developers were
> > notified about the availability of these patches in Ubuntu when the
> > transition began in Debian, though it looks like you chose not to
> > use it, and rebuilt the package instead.
>
> I wonder why I never received any bugreport about my stupid and wrong
> C++ transition here... After thinking now for the fourth time about
> this patch and my sort-of objection to the need I noticed that I did a
> C++ transition upload but forgot to change the library name
> appropriately. Though, exactly this is what I would have wished for:
> Why is this done in silent? Why not sent along?
Do you not read debian-devel-announce?
http://lists.debian.org/debian-devel-announce/2005/07/msg00001.html
> > In other words, I don't see what it is that you're dissatisfied about,
> > in your role as maintainer of these packages.
>
> I am dissatisfied about that noone informed me about my wrong C++
> transition upload, where there clearly seem to have been someone in
> ubuntu who noticed it.
The C++ transition for your package was completed in Ubuntu in 2005-06. The
availability of patches in Ubuntu for the C++ transition was announced on
debian-devel-announce in 2005-07. In 2005-08, you uploaded the package
during Debian's transition, but explicitly only rebuilt it rather than
renaming the package as required. How is Ubuntu responsible for this?
> > Are you speaking for yourself or on behalf of someone else?
>
> Mostly about myself, but given that I'm not the only one in this thread
> who complains about not getting any notifications about patches done
> within ubuntu that are relevant for Debian you can see it that I'm
> speaking for some others, too.
As far as I can see, your only first-hand experiences with Ubuntu are:
- Ubuntu redistributes your package, and you object to the fact that we do
so without having contacted you personally to notify you
- You were provided with a patch for the C++ transition, but declined to
apply it
I don't think that this experience justifies the animosity that you are
directing at Ubuntu.
--
- mdz
Reply to: