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

Re: Is vrms really still a Virtual Richard M. Stallman?

On Fri, Nov 21, 2003 at 10:42:12AM +0000, Henning Makholm wrote:
> Scripsit Anthony Towns <aj@azure.humbug.org.au>
> > On Wed, Nov 19, 2003 at 06:47:34PM +0000, Henning Makholm wrote:
> > > Me neither. A "virtual debian-legal" would be something that analyzed
> > > licenses:
> > Only if you assume a virtual foo does everything the regular foo does.
> No, but a virtual foo usually does *something* that the regular foo
> also does. Since debian-legal's primary task is to analyze licenses,
> that would be the natural expectation for a virtual d-l.

RMS's primary task is to advance free software, not rant at people
running non-free software too.

> Renaming vrms to a name derived from "debian-legal" would be as much
> an injustice to debian-legal as "vrms" is to RMS.

I'm not sure you can do an injustice to a mailing list.

> > > $ debian-legalint COPYRIGHT.foo
> > > COPYRIGHT.foo:33: warning: mentions specific protocol standard
> > > COPYRIGHT.foo:57: talks about "best efforts" to contact upstream
> > > COPYRIGHT.foo:64: US export control laws
> > $ debian-legalint realplay
> > Component: non-free
> > Limitations:
> >  no-source
> >  non-debian
> >  no-redistribution
> Apart from the syntactic details of the output format, how is this
> different?

It's trivial to write a program that tells you if an installed package is
non-free -- that's what vrms does. It's non-trivial to write a program
that tells you if the license in a text file is non-free.

The reason it's trivial is that we've already done the analysis,
and noted this down in the package's control file; it would be
straightforward to do the same thing for other interesting properties
of the package. Particularly things like "no-modify", "no-modify+dist",
"no-redist", "no-redist-for-profit", "no-commercial-use", "no-autobuild".

Being able to filter on these things would be useful -- it'd be nice
to be able to point aptitude at non-free, but have it automatically
ignore packages that you can't use because you're setting up a server
for a business.

Somewhat conversely, it might become valuable to have a "fsf-says-ok"
tag that can be stuck on packages that aren't DFSG-free, but that the
FSF thinks are free.

> > Something like that would probably be both useful and feasible,
> Feasible? Well, perhaps AI has made significant breakthroughs while I
> was looking the other way, but I doubt it.

Now, I was assuming that we'd pre-parse the licenses and have "legalint"
just go from data in a /var/lib/dpkg/info/ file of some sort, but it
occurs to me that in many cases, a fairly simple Bayesian analysis would
probably do a pretty good analysis of most non-free licenses.


Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.

Australian DMCA (the Digital Agenda Amendments) Under Review!
	-- http://azure.humbug.org.au/~aj/blog/copyright/digitalagenda

Attachment: pgpb3LWJc8KEm.pgp
Description: PGP signature

Reply to: