- From: David Booth <david@dbooth.org>
- Date: Tue, 20 Nov 2018 10:11:04 -0500
- To: public-lod@w3.org
On 11/19/18 8:19 AM, Hugh Glaser wrote: > It sounds to me like all you need is a URI that doesn't resolve (but could at a later date if you wanted to add meaning to it). While I agree with the advice given by Hugh (above) and others in this thread, I think Laura has hit on a deficiency that current RDF practices do not adequately fill. Blank nodes are second-class citizens in RDF. They cannot be used as stable identifiers in follow-up SPARQL queries, and as Laura pointed out, they cannot be used as predicates. And http URIs often place too much of a burden on the RDF producer, who may not be ready/able/willing to make the commitment to supporting a domain name and making those URIs usefully resolvable. Http URIs that will never usefully resolve do more harm than good. I will have more to say as part of a much larger topic in the next few days, but for the moment my suggestion to Laura would be in line with what others have suggested: use a non-resolving http URI if you think you can later make it usefully resolve; use a relative URI if it fits your use case; or use a URN. David Booth
Received on Tuesday, 20 November 2018 15:11:26 UTC