- From: Sean B. Palmer <sean@mysterylights.com>
- Date: Fri, 30 Nov 2001 22:36:37 -0000
- To: "Peter Crowther" <peter.crowther@networkinference.com>
- Cc: <www-rdf-interest@w3.org>
> That worries me. [...] How does this work when you > load fragments of RDF from two or more sources? Well, let's take an example:- [[[ _:x <#name> "Sean" . ]]] (there is something which has the name "Sean") merged with another document:- [[[ _:x <#name> "Peter" . ]]] (there is something which has the name "Peter") might produce something like:- [[[ _:x <#name> "Sean" . _:x1 <#name> "Peter" . ]]] It's really no big deal: if you have a collision, just rename the node and any subsequent nodes. I found it extraordinarily easy to implement, as I'm sure has anyone else that's written an NTriples parser. > Anyway, (at least half-seriously) why can't the authoring > program annotate the anonymous node with a triple > denoting the preferred label of the bNode within the > authoring environment? Why bother? Just take the current label to be the preferred label :-) There are plenty of different ways that one could approach the task, with varying levels of readability and processing to ensure that the labels remain consistent with the input, but the only real conditions are:- * Do not merge nodes that do not have the same label (in the document) * Do not merge any nodes in different documents, no matter what the label Both rules are a collorary to the fact that bNodes are scoped to the document (input) in question. Cheers, -- Kindest Regards, Sean B. Palmer @prefix : <http://webns.net/roughterms/> . :Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Friday, 30 November 2001 17:36:39 UTC