- From: Lin Clark <lin.w.clark@gmail.com>
- Date: Fri, 14 Oct 2011 11:28:24 +0100
- To: Philip Jägenstedt <philipj@opera.com>
- Cc: public-html-data-tf@w3.org, Richard Cyganiak <richard.cyganiak@deri.org>
- Message-ID: <CACho_As+StdFR=aVLxwxioEw7rsQt1HjrK_Bt5sbEoe2Nou-fg@mail.gmail.com>
On Fri, Oct 14, 2011 at 9:52 AM, Philip Jägenstedt <philipj@opera.com>wrote: > > > This kind of mapping can only really be done on a vocabulary-specific > basis, for vocabularies that define both forms to be equivalent. > There's an idea for how to deal with the vocabulary-specific nature of microdata's itemprops which came out of discussions Richard Cyganiak and I had a few months ago. Richard has given it more thought than I have, so I'm CCing him here. Hopefully he has a chance to give his input. There could be a registry that indicates whether or not the itemtype defines both the token and the URI form to be equivalent. This registry could also define a small number of patterns for forming the URI from the combination of the itemtype and the token. For example, one pattern would be to take the domain, without the path to the itemtype, and append the itemprop token as schema.org does. There would probably be around 3 patterns. Then, a vocabulary could be registered as following one of those patterns. By defining these patterns, it would give clear indication to vocabulary publishers how they can create a vocabulary that works for both microdata and RDF. Also, since the number of patterns would be limited to a small set with clear rules, it wouldn't put too much burden on consumers that want to translate microdata into RDF. Anyone would be able to identify in the registry which pattern a vocabulary uses (or if it does not follow a pattern because it does not allow URIs), so this information could be crowdsourced and wouldn't depend on the vocabulary producer registering it. I imagine it could be much like prefix.cc in its interaction. -Lin
Received on Friday, 14 October 2011 10:28:52 UTC