Re: DEP-5 meta: New co-driver; current issues
On Thu, Aug 12, 2010 at 01:47:42PM -0700, Manoj Srivastava wrote:
> > On to, 2010-08-12 at 14:59 +0200, Bernd Zeimetz wrote:
> >> On 08/12/2010 02:45 PM, Lars Wirzenius wrote:
> >> - Migrating all packages to the new format is an insane task which
> >> would take a *long* time and a lot of work.
> > There is no goal to migrate all packages. That's a strawman.
> I found that surprising; perhaps I have forgotten a lot about
> this proposal. So, if I understand this correctly, this proposal is to
> come up with some way of creating a standard format for copyrights that
> is not meant to be universal (since lacunae that make it unusable for
> some packages are not going to be addressed), and not all packages will
> ever have a machine readable copyright?
The goal of this DEP is to craft a machine-readable copyright format that
has broad support from the Debian community and is *suitable* for use in any
That's quite a different goal from proposing that it be *adopted* in all
packages, or mandating such adoption.
Many of the objections that have been raised on this mailing list to DEP-5
are objections to the very idea of a machine-parseable format. No
machine-parseable format is ever going to have zero syntactic overhead, nor
will it satisfy developers who are opposed to the idea of a structured
debian/copyright. It is not a productive use of anyone's time to engage in
discussions on this mailing list about the format being "heavy" or "ugly" or
"too much work" when those advancing these claims offer no suggestions for
improvement. There are almost 3000 source packages in Debian today that are
already using some variation on DEP-5. When the rate of adoption is this
high for something that's still a half-baked draft, arguing about the /need/
for a machine-parseable format, as some have continued to do, is missing the
point. That avalanche has already started, it's too late for the pebbles to
Instead, the task at hand is to craft a file format that does a good a job
as possible of meeting the disparate demands placed on debian/copyright, so
that as maintainers *do* implement it, we get something that's useful to
Debian instead of something that's a burden. To that end, concrete
suggestions for improving the format are definitely welcome.
> I had hoped that we would try for a format simple enough to be
> generally usable (just a header Type: GPL; Subtype: V3) would address a
> lot of use cases
Splitting the license name and version number into separate fields implies
that one can do useful analysis using only the license name. I don't
believe this is the case; I don't know of any license family whose versions
are bidirectionally compatible, and as for third-party license auditing, the
most common free license that people want to avoid nowadays is GPLv3 because
of terms not present in GPLv2.
> without being onerous and much more detailed than what is required now.
> Do we know what use cases we are trying to address? Just having
> a machine parseable list of licenses per package (perhaps with known
> tags for popular licenses, and other) seems to be far more simply
> addressed than the complex schema I seem to recall being bandied
The complexity of the schema is simply a reflection of the richness of the
information that maintainers already ship in their debian/copyright files
today. Listing separate licenses for separate source files is already a
common practice; DEP5 is just allowing maintainers to represent this
information in a structured manner.
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/