On Wed, 5 Jan 2005, Pete Johnston wrote: >> On the other hand, the problem with using >> a very generic link type like 'meta' is that robots have to visit the >> linked metadata before they can determine if it is useful to >> them (i.e. >> they might find LOM/XML, they might find DC/XML or they might find >> something else at the end of the link). > > The proposal also recommends including a LOM/XML-specific MIME type > > <link rel="ieeelommetadata" type="application/ieeelom+xml" title="LOM > Metadata" href="lom.xml" /> > > and says they both "shall" be used. > > So the crawler could use that MIME info to decide whether it wants to > dereference the target URI. I don't see why it needs _both_ the > LOM-specific rel attribute value and the LOM/XML specific MIME type. Duh... sorry. Yes you're absolutely right. (It'd help if I actually read the thing before commenting!). I note that application/ieeelom+xml doesn't appear to be a registered MIME type. Is it being registered? (I guess that we should similarly think about registering something like application/dc+xml for the non-RDF XML encoding of Dublin Core - the RDF encodings of DC can use application/rdf+xml, which is registered?). Andy -- Distributed Systems, UKOLN, University of Bath, Bath, BA2 7AY, UK http://www.ukoln.ac.uk/ukoln/staff/a.powell tel: +44 1225 383933 msn: [log in to unmask] Resource Discovery Network http://www.rdn.ac.uk/