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

Re: census: missing source packages

On Sun, 19 Jun 2011 14:41:43 +0800
Paul Wise <pabs@debian.org> wrote:

> I've been looking at the apt sources.list snippets included in the
> census and despite my prodding late last month, there are still five
> distributions that do not provide source packages despite offering
> binary packages. In one case I've even found users complaining on their
> distribution's forums about this. I've verified that in each of the
> cases where I can't find source packages, there are LGPL or GPL binary
> packages listed in the derivative's Packages files. So not only are some
> of our derivatives not living up to FLOSS principles, but a few are
> likely even violating licenses and copyright law.
> Please note that I only checked the derivatives with a sources.list
> snippet and with no deb-src lines in their sources.list. I still need to
> check those distributions who don't have sources.list snippets in the
> census and for the ones with deb-src lines, check that the deb-src lines
> match the deb lines.
> Now, my question is; what should Debian do about this if anything?

As Debian cannot actually stop anyone distributing Debian packages
themselves, it may have to come down to "naming-and-shaming" once we
are certain of the situation for each one. If users are already
complaining, without results, then adding a complaint from Debian might
not add much.

However, at least if such derivatives are identified *on Debian pages*
as "questionable", we can be seen to be putting our users first and
making our position clear. It would be better if derivatives declared
this themselves (it may prompt action), so this should be part of the
census. Let's make it an explicit question:

"Does the derivative provide full source for all packages where the
licence of the package requires the distribution of source alongside

In the end, if the derivative becomes less popular compared to similar
derivatives which do provide source, that may be more useful in the
wider scheme of things.

It's part of the Social Contract, as I see it. Where Debian provides
information about a derivative, that information should include some of
the negative aspects of that derivative where the derivative is not
upholding concepts which users would expect from other Debian
derivatives. Especially where those aspects impact directly on the
Social Contract and/or DFSG.

In some ways, the current census misses this kind of data. There might
need to be section(s) along the lines of (links to webpages would be

"Why use this derivative?"
"What are the differences between this derivative and standard Debian?"
"What do users not get from the derivative which Debian provides?"

e.g. for Emdebian, we try to cover the first question on our index page:
Why use Emdebian rather than Debian?
The most obvious reason is installation sizes. Debian simply cannot fit
onto some devices that could run GNU/Linux. Other machines can
accommodate a typical Debian installation but would have little
available space for user data etc. and adding more storage is
impossible or impractical.

Other answers could be:
manpages, info pages, documentation and changelogs are removed.
Translations are moved out into separate binary packages, selectable by
locale and by package.

What's missing?
Not all packages are converted for Emdebian, either for the Grip or
Crush distributions. Tools are provided (within Debian) to convert
packages to Grip if not already provided. Not all architectures
available from Debian are supported.

Maybe the list of supported architectures should also become a census

Indeed, there's possibly scope for the census to include more free text
and be less of a multiple-choice tick-the-box list.


Neil Williams

Attachment: pgpDmCpRoDU4n.pgp
Description: PGP signature

Reply to: