Elias Torres

Hi. Follow me on Twitter.

1. libby 03.19.06 / 6am
   Elias, you might want to ask on the foaf mailing list, because some of these questions are a matter of opinion: http://rdfweb.org/mailman/listinfo/rdfweb-dev
   
   My take:
   
   1. I’d go for one foaf file for each person. Link them with the uris for the person and a rdfs:seeAlso to their foaf file and (preferably) some IFP like mbox or mbox_sha1sum. This makes prtimaryTopic and the like easier.
   
   2. Having several and linking them is fine, IMO.
   
   3. Filling in an example Foaf-a-matic http://www.ldodds.com/foaf/foaf-a-matic.html is a good way to find out the best form for primaryTopic etc.
   
   Hope that helps
2. Morten Frederiksen 03.19.06 / 6am
   1: As always: It depends. It’s the basic organizing question, whether to keep one big store or several small ones.
   
   2: Links (seeAlso) are always good, it’s the web thing…
   
   3: For a file describing only an individual, the primaryTopic should be just that. For a file describing more than one person, there really isn’t a primaryTopic, use multiple topic’s instead.
3. Reto 03.19.06 / 1pm
   Don’t do it, don’t give yourself a URI, use IFP instead. Nothing makes you unique but you . My suggestion would be to make a derefereceable resource with your family as primary topic, and one for every family member.
4. Elias 03.19.06 / 3pm
   Thanks Libby, Morten and Reto. I think I’ve got it right. I have a family foaf and an individual foaf. I used the multiple foaf:topic tags to point to everyone in the family. I’m using content negotation. I’m using /who instead of /foaf to not be attached to the foaf prefix. What else? I’m exhausted after reading the entire thread from TimBL and all of the comments. Of course the first comment is Reto’s going against URIs. I think Reto might like LSIDs.
   
   I’ll never win, but at least I use both IFPs and URIs, let’s hope for the best. Anyways, I’ll have a trust in my name that maintains my domain forever. But then again I won’t be needing a URI forever.
5. Lee Feigenbaum 03.19.06 / 11pm
   Well, I’m late to the game, but since Elias called me out by name I’ll mix metaphors and throw my two cents into the ring.
   
   1) While (as others have said) it doesn’t make a difference in theory, I prefer having a single FOAF document. The convention of using rdfs:seeAlso to point to other FOAF files is just that–a convention. The seeAlso predicate doesn’t contain any more substantial semantics than does a regular ol’ hyperlink in the (non-semantic) web.
   
   2) Perhaps foaf:primaryTopicOf could be used instead of rdfs:seeAlso? Then you could have four documents for your (current) family, each with an unambiguous foaf:primaryTopic. They could each reference (via foaf:knows, for instance) the others and use foaf:primaryTopicOf (rather than rdfs:seeAlso) to point to the others. This sounds like a pain to maintain though…
   
   I haven’t done my homework: is foaf:primaryTopicOf (inverse of foaf:primaryTopic) part of the foaf spec?
   
   3) you seem to have this part well figured out.
   
   Wanted to point out that the OWL reasoner inside my head currently is concluding that Alejandra, Briana, Noah, and you are all the same person, given your shared foaf:mbox_sha1sum value
   
   About URIs: I like giving yourself (and others) URIs in addition to liberal use of IFPs. As Reto points out in his comment on TBL’s post, aggregators need to be aware of both IFPs and also multiple URIs for the same foaf:Person. That’s life, but I don’t see it as a reason to shy away from URIs. The possible presence of multiple URIs per person and of IFPs does make SPARQling (in the absence of a reasoner) a bit harder in the calendar demo (which also does a bit of multi-URI and IFP smushing on foaf:People) — I’ll write a bit about that later this week, hopefully.
   
   Lee
6. Dan Smith 03.20.06 / 3am
   Looking at it through my LSID+caching authority blinders/bias, I think using a CDB representation is the way to go.
   
   http://www.w3.org/Submission/CBD/
   “a body of knowledge about that resource which does not include any explicit knowledge about any other resource which can be obtained separately from the same source.”
7. Dan Smith 03.20.06 / 4am
   CBD.. spelling
8. Elias 03.20.06 / 8am
   Lee, stupid Vi swp files, I had fixed that, but somehow I ended up with same mbox_sha1sum values again. Fixed. Tx.

Powered by Hemingway flavored Wordpress.

Entries RSS Comments RSS