On Thu, Jan 5, 2012 at 11:05 AM, Ross Gardler <[email protected]> wrote: > On 5 January 2012 13:57, Rob Weir <[email protected]> wrote: >> 2012/1/5 Jürgen Schmidt <[email protected]>: >>> On 1/5/12 12:12 PM, Ross Gardler wrote: >>>> > > ... > >>> And the proposed index only solution is not satisfying for me because of the >>> explained reasons (missing hosting of the binary extension packages or >>> templates). But I would like to learn more about it and especially how it >>> should be maintained, the frontend how users would register their extensions >>> ... >> >> Let's take an example of the general pattern, with blogs. ... they all >> agree >> on some basic standards: HTML for the content, and RSS or Atom for >> syndication. > > ... > >> So the model that I think is attractive for extensions and templates is: > > ... > >> 2) What we do specify are: >> >> a) The metadata related to extensions. > > ... > >> >> b) An encoding of the metadata, in XML. > > At the risk of diverting this thread too far from the goal of coming > up with an overall plan for moving forwards short-medium-long term... > > Consider the RDF Description of a Project format. The ASF uses it for > http://projects.apache.org That site is generated from a bunch of > PERL, XSL and the like pulling content from wherever the projects want > to host their DOAP descriptor and publishing a catalogue. It's not as > feature rich as needed for AOO but it is a good proof of concept. > > For an intro to DOAP see http://www.oss-watch.ac.uk/resources/doap.xml > > Remember that there are a whole load of ASF projects dealing with RDF > data, it's quite possible some of those will be keen to help us if > this were a route we took. >
RDF is fine as well. The general thought is not to reinvent the wheel with our own ad-hoc format. Pick something that has broad support in tools and libraries already. We're fortunate to have several good options for this kind of thing > Ross
