(we should maybe continue this thread on soc-coordination only? Keeping Olivier in Cc anyway) Hi Olivier, Quoting Olivier Berger (2013-04-10 13:35:13) > I think that it would be great, for maximum interoperability (derivative > distros, upstreams, others), if the data about Debian artifacts could be > using semantic formats such as the one I've tried to add to the PTS . > > So in addition to offering anyone (inside or outside Debian) the > possibility to interface with the events happening in that > infrastructure, it would convey non-ambiguous semantics out of the box. > > So, instead of mentioning an upload about "apache2" (a string that is > supposed to deal represent a source package of the Apache 2 HTTP > server), a message could reference > <http://packages.qa.debian.org/apache2#project> which is a > dereferenceable URI (see http://packages.qa.debian.org/a/apache2.ttl) > where meta-data about that source package is available. Thanks for your input and the URLs, I didn't know about those. While I like the idea of semantics in the data, if I receive a message from dak saying that apache2 has been uploaded, I don't expect it to contain anything else than the name of the package for the simple reason that it's pretty much all dak knows about it at the time of the upload. Since we have a simple URL scheme for the Turtle data, I would simply let the listener use the package name to fetch the data if they need it. However, if this data is generated — I might be very wrong on that, I have no clue at all — you could simply send a message when the generation is done, which might be of bigger interest for those using the semantics than the information about the upload. To me, this infrastructure should be lightweight, since it would be impossible to give out URLs fitting all intents and purposes.