[darcs-users] RDF metadata for patch files

[Max Battcher]

Trent W. Buck wrote:
> I'd like to propose (long term) adopting either RDF or some equivalent
> that allows end users to say on a per-patch file "I wrote this work, I
> have copyright on it, and I license you to use it per the following
> contract (e.g. GPL-2+)".
> 
> There would be facility so that an end user can declare their preferred
> license, and a repository can declare its preferred license, and Darcs
> would warn the user if they were incompatible.  For example, if I want
> to use CDDL for my works, but the Darcs repository is GPL-2'd, it would
> complain when I tried to do a "darcs record" unless I added to the
> command line --license gpl-2+ or whatever.
> 
> http://en.wikipedia.org/wiki/Resource_Description_Framework

RDF is a "standard", but it's a long rabbit hole to fall into. License 
data may be useful to encode in a repository, but I'm not entirely sure 
if RDF is the way forward...  On the other hand RDF is a hell of a 
hammer and you could make the case that a lot of other nails that darcs 
should solve could be solved with rich RDF support.

The user interface alone will be a lot of work, but just thing of all 
the many other things we could discuss (and this is just of the top of 
my head; with a little Wikipedia skimming):

Should we push XML or Notation 3? Notation 3 for CLI, XML for 
--xml-output? XML for both?

Should there be SPARQL support?

Would a triplestore be necessary/helpful?

Should darcs constrain to prescribed ontologies or allow for free-form RDF?

Should the RDF graph just involve Patches? What about per-file, 
per-repository metadata? Perhaps each repository might have a single 
triplestore and we can develop/re-use ontologies that let us store 
per-contributor knowledge (email, FOAF, DOAC, favorite color), per-file 
knowledge (license, mime-type), and anything else we might ever need. We 
can then work on intelligently commuting RDF graph changes...  Or not.

RDF is a "Semantic Web" technology that I'm still dubious as to the 
merits of general RDF use.

--
--Max Battcher--
http://worldmaker.net