- From: Tom Heath <tom.heath@talis.com>
- Date: Fri, 20 Mar 2009 19:20:18 +0000
- To: Knud Hinnerk Möller <knud.moeller@deri.org>
- Cc: public-lod@w3.org
Hi Knud, Just some short specifics in response... 2009/3/19 Knud Hinnerk Möller <knud.moeller@deri.org>: > Hi again, > > On 19.03.2009, at 14:58, Tom Heath wrote: > >> Hi Knud :) >> >> 2009/3/19 Knud Hinnerk Möller <knud.moeller@deri.org>: >>> >>> Hi Tom, >>> >>> great stuff, and nice you're using some of our swc vocabulary! >> >> Once a <http://data.semanticweb.org/ns/swc/ontology#DogfoodTsar> >> always a <http://data.semanticweb.org/ns/swc/ontology#DogfoodTsar> ;) > > It's an unliftable curse! No comment ;) >>> I'd love to >>> import the data into the dog food site - the more data we have there, the >>> better. However, as your data is at the moment, only the people would >>> show >>> up as resources on our site, since everything else is in the ldow >>> namespace. >> >> Good point! >> >>> Do you think you could generate an xml file as described in [1]? That >>> would >>> make the import task easiest and would generate RDF that integrates >>> nicely >>> with the rest of the dataset! >> >> No sign of [1] in your mail, but I think I follow you. So am I ok to >> modify my script to mint URIs in the following namespace? >> http://data.semanticweb.org/workshop/ldow/2009/ > > ooops... [1] was supposed to point to > http://data.semanticweb.org/documentation/user/faq#how_to_add > > You are absolutely ok to mint URIs in the dogfood namespace, as long as we > manage to get the data into the site afterwards! The easiest way might be to > produce the XML files as outlined in [1] (this time the link is really > there! ;), but we should also manage to get it to work by adding the right > triples in your RDFa. I should write up a list of triples that are necessary > for the dogfood UI to make sense of the data - everything can be loaded, but > not everything displays nicely at the moment. > >>> One comment about the modelling. You say: >>> >>> <http://events.linkeddata.org/ldow2009/papers/ldow2009_paper11.pdf> a >>> swrc:InProceedings . >>> >>> Would it not be nicer to have a separate URI for the paper and then link >>> to >>> the pdf? E.g.: >>> >>> <http://events.linkeddata.org/ldow2009/papers/ldow2009_paper11> a >>> swrc:InProceedings . >>> <http://events.linkeddata.org/ldow2009/papers/ldow2009_paper11> swrc:url >>> <http://events.linkeddata.org/ldow2009/papers/ldow2009_paper11.pdf> . >>> >>> What do you think? >> >> Yes, good point, this would make good sense. I'll make these >> improvements when I have a spare moment. > > Very cool. > >> Before I spend too long on >> that though, are there plans that you or anyone else knows about to >> RDFize the data about all WWW2009 workshops etc? Would be a shame to >> duplicate effort, though providing it at source does have a lot of >> appeal. > > I just asked Daniel Schwabe if he knew if we could repeat the process from > last year (take EasyChair data, convert with our script) for the main > conference data. Let's see what comes out of that. I don't know anything > about the workshops, though. I'd say there are three possibilities at the > moment (in increasing order of effort involved): > > 1.) If you use EasyChair, take the data from there and convert with existing > script. You need access to the XML output of EasyChair, though. > 2.) send me the data in XML as per [1], so that I can convert it with the > same script. > 3.) generate the RDF yourself. That will probably need some iterations to > get it right (or I will have to make the dogfood URI more robust!) 3. makes the most sense in this context, as we don't have access to the EasyChair XML output and 2 would duplicate effort when 3 is already in place. >> Incidentally, I think there are some errors in >> <http://data.semanticweb.org/workshop/LDOW/2008/rdf>. AFAICR there are >> no hasAcronym and completeGraph properties in SWC. > > ouch, you spotted my vocabulary hacking there... those two properties are > additions I came up with, which are unfortunately not yet reflected in the > vocabulary source online. I'll correct that asap. Aren't these both fairly generic properties that may best live somewhere less conference-specific? Picking up on Damian's point, I'm also not a huge fan of swc:url. In the past did we not just do sameAs between the 'uri for the paper' and the uri for the corresponding pdf, giving a bit of a handwave to any frbr-like issues? Having said that, frbr:realization would seem a better way to proceed long-term, as Damian suggests. The curse lives on ;) Have a nice weekend, and enjoy eating something else other than DogFood. I'd recommend a good pizza. Tom. -- Dr Tom Heath Researcher Platform Division Talis Information Ltd T: 0870 400 5000 W: http://www.talis.com/
Received on Friday, 20 March 2009 19:21:02 UTC