On 9/22/07, Stefano Mazzocchi <[EMAIL PROTECTED]> wrote: > Matthieu Riou wrote: > > Hi, > > > > So far in Discordia, to build the license database, a script browse the > > central Maven repository and extracts license information from project POMs. > > So we can at least do a best effort to initialize the database with > > hopefully accurate data (btw if somebody now of other formal sources that we > > can use to know the license of a given project/artifact, let me kow). The > > license information for each artifact (read jar, war, mar, gem, whatever) is > > then serialized in a file, on file per artifact (with a well chosen name > > that facilitates indexing). > > > > So far the format used for this license file is as simple as it gets: > > > > in org.springframework-spring-jms-2.0.1.xml > > <artifact> > > <name>spring-jms</name> > > <project>org.springframework</project> > > <version>2.0.1</version> > > <licenses> > > <license> > > <name>The Apache Software License, Version 2.0</name> > > <url>http://www.apache.org/licenses/LICENSE-2.0.txt</url> > > <distribution>repo</distribution> > > </license> > > </licenses> > > </artifact> > > > > I think there are more pertinent RDF-based formats around. Specifically, I > > think that Robert came up with something like this for a similar > > application. Maybe we could reuse the same thing?
i was just playing around with one of http://simile.mit.edu/wiki/RDFizers > The DOAP ontology has licensing information. > > http://usefulinc.com/doap/ > > which is also used by O'Reilly CodeZoo > > http://www.codezoo.com/ IMHO there are two sets of data that need to be collected and maintained. (not sure whether both are in scope for discordia or whether a separate mini-project might be a better home for the license meta-data.) the first links artifacts to licenses. the DOAP vocabulary should work fine for this. (indeed, DOAP would be another good source for meta-data.) probably worth posting a hello to the DOAP mailing list. the second is meta-data about licenses. so, it's not enough to know that groovy has a particular license but also that this license is from the apache license 1.1 family. licenses in the apache 1.1 family share the same legal properties: they are open source, they require attribution, they are non-reciprical and so on. they also all share the same apache policy WRT inclusion as dependencies. this seems to me quite a new and different vocabulary to DOAP. might be a good idea to talk to the DOAP people about this new vocabulary. i think that the best approach would be not to normalize the license URLs (DOAP has a small set of predefined URLs) but take them straight from the source. the license meta-data would then link URLs to license family and so to their general characteristics. - robert --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
