- From: Chris Bizer <chris@bizer.de>
- Date: Sat, 4 Sep 2010 22:01:35 +0200
- To: "'Dan Brickley'" <danbri@danbri.org>, "'Anja Jentzsch'" <anja@anjeve.de>
- Cc: <public-lod@w3.org>, "'Thomas Baker'" <tbaker@tbaker.de>
Hi Dan, > > This is great! Glad to see this being updated :) > > One thing I would love in the next revision is for FOAF to also be > presented as a vocabulary, rather than as if it were itself a distinct > dataset. While there are databases that expose as FOAF (LiveJournal > etc.), and also a reasonable number of independently published 'FOAF > files', the technical core of FOAF is really the vocabulary and the > habit of linking things together. Having a FOAF 'blob' is great and > all, but it doesn't help people understand that FOAF is used as a > vocabulary by various of the other blobs too. Yes, we also felt that having a blob is a bit misleading and were thus thinking about using a cloud icon for FOAF and SIOC to reflect the fact that the blob actually consists of many separate files on many different servers. Beside, we have started to tag datsets in CKAN with the vocabularies that they use. So, ideally all datasets that use FOAF should be tagged with format-foaf and people can use this data via the CKAN API to draw any visualization of the LOD cloud they like. > And beyond FOAF, I'm > wondering how we can visually represent the use of eg. Music Ontology, > or Dublin Core, or Creative Commons vocabularies across different > regions of the cloud. Maybe (later :) someone could make a view where > each blob is a pie-chart showing which vocabularies it uses? Interesting idea. I would also love to see this. Maybe we can give it a try, otherwise of course everybody is invited to get the data from the CKAN API and visualize it in any way he thinks is interesting. > As a vocabulary manager, it is pretty hard to understand the costs and > benefits of possible changes to a widely deployed RDF vocabulary. I'm > sure I'm not alone in this; Tom (cc:'d) I expect would vouch the same > regarding the Dublin Core terms. So if there could be some view of the > new cloud diagram that showed us which blobs (er, datasets) used which > vocabulary (and which terms), that would be really wonderful. On the > Dublin Core side, it would be fascinating to see which datasets are > using http://purl.org/dc/elements/1.1/ and which are using > http://purl.org/dc/terms/ (and which are using both). Similarly with > FOAF, I'd like to understand common deployment patterns better. I > expect other vocab managers and dataset publishersare in a similar > situation, and would appreciate a map of the wider territory, so they > know how to fit in with trends and conventions, or what missing pieces > of vocabulary might need more work... Yes, having data about usage patterns would be great. But I guess there are also limits to the meta-data that people can gather manually. So the best would be if somebody would run a crawler and extract meta-data about vocabulary usage and other usage pattern directly from the LOD datasets. Nobody has done this yet but hopefully somebody will soon start doing this. Cheers, Chris > Thanks for any thoughts, > > Dan
Received on Saturday, 4 September 2010 20:02:08 UTC