Sometimes we get so excited about using URIs for everything that we use handy URIs like email addresses and homepages to identify people, ala
<baseballNotes> dc:creator <mailto:bob@example.org>. <http://www.purl.org/net/danbri/> a :Brit.
Don't do that. why not? For one thing, it causes people to get confused about how RDF works; but real reasons should go here.
Instead, use a property to relate the person to their mailbox or homepage:
<baseballNotes> dc:creator [ foaf:mbox <mailto:bob@example.org>]. [ foaf:homepage <http://www.purl.org/net/danbri/> ] a :Brit.
Also consider foaf:page, a superproperty of foaf:homepage, and its owl:inverse, foaf:topic.
<http://www.purl.org/net/danbri/> foaf:topic [ a :Brit ].
In TopicMaps this is called a subject indicator. Let's give it a home in URI space... hmm...
@prefix tm: <@@hmm...>. tm:subjectIndicator a r:Property; s:label "subject indicator"; = foaf:page; owl:inverseOf foaf:topic, cyc:containsInformationAbout-focally; s:isDefinedBy <@@where's the topicmap spec?>.
As to the related subject address notion in TopicMaps, that's just identity, i.e. owl:sameAs.
tm:subjectAddress owl:samePropertyAs owl:sameAs.
Discussion...
Curing the Web's Identity Crisis Steve Pepper, Ontopia (Mar 2003). chumped/discussed 3Apr, wherein DanC noted this article is confused about resources vs. representations, for one thing, and about how hashes work, for another
discussion 10Apr among StevePepper, EricMiller, etc. previous rdfig chat, Living with topic maps and RDF
Disambiguating RDF Identifiers by sandro suggests that RDF could be changed to support URIs being used both directly and indirectly, but this proposal found little support in the core RDF community. He also found it challenging to implement efficiently.
For a while, DanConnolly was using dc:subject ala foaf:topic. But inquiries failed to confirm that this is per community consensus. dc:subject is evidently a relationship between a document and some subject keywords, perhaps taken from a controlled vocabulary. Hmm... ThingsVersusTheirNames.
- is deployment of dc:creator with string values so widely deployed that we should just change the schema to say that it gives a name for a creator?