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

Re: What are the benefits of a machine-parseable ‘debian/copyright’ file?

On Mon, 23 Mar 2009 10:03:10 +1100
Ben Finney <ben+debian@benfinney.id.au> wrote:

> * Display of the copyright notices, license grant and terms (in e.g.
>   ‘aptitude’) when deciding whether or not to install a package.

I don't see how that is a real gain, except for packages in non-free.

Are we looking for a problem to justify the solution perhaps?

Are you saying that you don't want any BSD packages or that you're
trying to make Debian out of only packages using GNU licences? This
thread gets more and more insane.

(I thought we were approaching some degree of sanity earlier on, but
that has been left behind in this sub-thread, which is apparently
trying to invent illusory benefits to justify what is increasingly
sounding like a personal obsession.)

> * Display (or via a link) of the copyright notices, license grant and
>   terms on the package-specific pages at ‘packages{,.qa}.debian.org’.

Hasn't it been established that lists of names and email addresses of
copyright holders are not a priority for content because updating the
list imposes an intolerable workload on many maintainers, either with
lots of packages, large packages or packages with high upstream

debian/copyright should be restricted to only that data that is
essential for the LICENCES specified in the package and not all
licences require any names of copyright holders, let alone all. Let's
wait for the review of just what is required before trying to justify
the proposal with an illusory benefit that could be completely
undermined by the licences themselves.

> Perhaps you have never factored license terms into your decisions as a
> prospective user of a package, pre-installation; I do, all the time,
> since I want to know what I'm getting myself in for before adding a
> work to my computer's operating system.

The whole point of main is that this doesn't need to be a day-to-day
decision. Licences of packages in main only become an issue when
developing code that will link against it. There is no need for any
new tool for that purpose.

> As for copyright notices, if the work is in a field of interest to me,
> then I'm interested to know who wrote it. Reputation is one of the
> greatest benefits reaped by a developer of free software; having their
> copyright notice automatically and visibly connected with the package
> can only increase this, and I consider that a very good thing.

Which copyright holder? The first one or #543 ? Do you really care if
someone you know contributed one significant patch to a codebase of
tens of millions of lines? Do you really want an unverifiable,
unmaintainable and almost-guaranteed out-of-date list? For what
possible purpose? I see no gain here.

Have you watched the GNOME About box? It goes on for what seems
like weeks.

> > If I'm developing software and I want to use another piece of
> > software (e.g. library, framework or component) I check the license
> > (I don't care about the copyright holders btw.) and perhaps
> > sometimes I check the copyright file but most of the time I just
> > check what upstream put on their website. For these seldom uses a
> > common format or tools wouldn't help me much.
> The copyright notices and license terms on the work in Debian only
> infrequently bears relation to what upstream puts on their website,
> not least because a great many upstream websites have *no* copyright
> or license information easily discoverable.

Developers choose a new library, framework or component on a very
infrequent basis, even upstream. The licence is only part of the
decision process of such a change. It can take years to port an
application from one version of a framework to the next (witness
gtk1.2) let alone to a completely new framework. A few extra seconds to
look at the actual licence from the actual source is hardly a
significant loss - you'll be spending a lot of time reading the
source anyway.

> Then there is the (admittedly less frequent, but undiscoverable
> without looking at the Debian copyright information for the work)
> issue of the upstream copyright information differing from what was
> packaged for Debian.

Which only becomes more likely if you expect a list of copyright
holders in debian/copyright because it can never be 100% complete
across all of Debian, if only because upstream don't have a full list

> > I can understand there may be benefits of a parsable format but I
> > don't directly see enough gain. On the other hand there seems to be
> > a lot of (perceived) cost involved (maintainer work).
> As has been pointed out many times, there should be no significant
> increase in the amount of writing to generate one of the
> machine-parseable ‘debian/copyright’ files.

Only if the file does *not* require a detailed list of copyright
holders. Conversion of existing files is also a huge burden.

> It merely imposes a
> consistent structure, much like the ‘debian/control’ file, or the
> pseudo-header of a Debian BTS report.

Nothing about the proposal is "merely" anything.


Neil Williams

Attachment: pgpwuHmdzceE6.pgp
Description: PGP signature

Reply to: