- From: Dan Brickley <danbri@danbri.org>
- Date: Mon, 1 Nov 2010 09:14:33 +0100
- To: Karen Coyle <kcoyle@kcoyle.net>
- Cc: public-lld@w3.org
On Mon, Nov 1, 2010 at 1:01 AM, Karen Coyle <kcoyle@kcoyle.net> wrote: > Quoting Dan Brickley <danbri@danbri.org>: > > >> >> Is the distinction one concerning the 'things in the world', or more >> about their actual descriptions in some particular record? > > > Dan, this is one of those areas where the library cataloging view is very > particular but also very different from the SemWeb view. First, I suggest > taking a look at the diagrams that Gordon pointed us to: > > http://www.gordondunsire.com/pubs/docs/frdiagrams.pdf > > You will see there that names of things are first class objects in the > library world. The reasons for this are historical (not hysterical)[..] Thanks for confirming my reading! I think the big difference w/ the SemWeb world isn't so much around names, but around the idea of these standards all being used alongside each other in opportunistic mix-and-match combinations. That goes against many people's ideas of what a data standard amounts to, not just in the library world. Ok so on this topic of names, I long wanted a class in FOAF that stands explicitly for someone's name. The reasons for this I hope are also not hysterical, just pragmatic. Sometimes people have several names, and unless those are modeled as 'things in themselves', it is hard to keep their properties clearly separate. In particular I'd like to be able to add pronunciation information (using media files, and/or http://en.wikipedia.org/wiki/International_Phonetic_Alphabet) and possibly dates. (Genealogy apps might also want to treat names as first class things with information about their meaning, associated geography etc.). So why doesn't FOAF have a class whose instances are human names? (Just for now let's set aside Groups, Organizations, Projects and other Agents just as pedigree pets! all these deserve names too). The answer might be hysterical, it's certainly silly. Why? we just couldn't think of a good name for it, or for the main person-to-name relation. We already had 'foaf:name' that links a person to a textual form of the name, so we couldn't use that. If there was a class let's say 'foaf:NameDetail' (to avoid a class and property's id differing only in capitalization), and a linking property e.g. 'foaf:nameInfo', then any person could be linked tidily to multiple descriptions of names, and each of those might have dates (eg. a post-marriage name, or a post-moving-to-USA name), pronunciation, alternate spelling etc. I'm not sure NameDetail and nameInfo quite work as (er...) names, but maybe they're bearable. BTW as discussed here recently, SKOS has some related facilities; in fact SKOS's labelling machinery is useful as a general purpose 'escape hatch' wherever the RDF vocabulary world hasn't modelled some area in detail. I hope foaf:focus helps people exploit it this way. But that said, I do think names are in fact worth modelling explicitly. They're just so hard to do right (internationally etc.), that for years re FOAF it was too intimidating a problem to address at all (see fascinating links below). So - we relied mostly on a single flat text/markup string, foaf:name, plus half-hearted variants on familyName and givenName. But perhaps adding a *class* for a name is a nice next step as it would allow other parties to attach their own properties while sharing some common structure? cheers, Dan ps. via http://www.delicious.com/danbri/names http://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/ http://blog.jclark.com/2007/12/thai-personal-names.html http://rishida.net/blog/?p=100
Received on Monday, 1 November 2010 08:15:06 UTC