- From: Eric van der Vlist <vdv@dyomedea.com>
- Date: Tue, 08 Aug 2006 23:36:13 +0200
- To: Dan Connolly <connolly@w3.org>
- Cc: Bernard Vatant <bernard.vatant@mondeca.com>, semantic-web@w3.org, Franck Cotton <franck.cotton@insee.fr>
- Message-Id: <1155072974.26048.17.camel@localhost>
Le mardi 08 août 2006 à 15:55 -0500, Dan Connolly a écrit : > On Tue, 2006-08-08 at 22:37 +0200, Eric van der Vlist wrote: > > Le mardi 08 août 2006 à 14:09 -0500, Dan Connolly a écrit : > > > On Tue, 2006-08-08 at 19:02 +0200, Bernard Vatant wrote: > > > > > > > Note that what the tabulator will do is follow both seeAlso links > > > > > and merge the data from DOC1 and DOC2. That seems consistent > > > > > with what you intend: "equally normative". > > > > > > > > > > > > > > Yes. But it might happen that the merged data are not consistent, > > > > although of course separately DOC1 and DOC2 should be. > > > > > > That's not what "equally normative" means to me; "equally normative" > > > implies consistent, no? > > > > We might have two different issues. > > > > Short term, yes DOC1 and DOC2 will give complementary and consistant > > descriptions of the same set of entities. > > Ah... good. > > > Longer term, if we want to publish several versions of these data we > > will have to figure out how we would like to express all that. > > Yes, that's an interesting puzzle. Yes. I haven't had the opportunity to think about it in RDF world, but I have had to cope with this type of history issues in a prior life and they become rapidly very complex. > > Including > > rdfs:seeAlso references to different versions of the same data could > > lead to include inconsistent descriptions and I think that this is what > > Bernard means. > > In that case, I hope we can use something more refined than seeAlso > to say "well, here's the newer version, i.e. the version I'm most > actively supporting today, though I still stand by this old version. > I don't promise you can use them together at the same time to > get something sensible." and maybe the tabulator can be taught > to discriminate. Yes, something similar to what OWL uses for its own purpose (owl:priorVersion) would be nice. It would also be interesting to attach a date to a data set. This seems to be a rather generic issue. Are you aware of anything that we could borrow for this kind of issues? Thanks, Eric > > > Eric -- GPG-PGP: 2A528005 Le premier annuaire des apiculteurs 100% XML! http://apiculteurs.info/ ------------------------------------------------------------------------ Eric van der Vlist http://xmlfr.org http://dyomedea.com (ISO) RELAX NG ISBN:0-596-00421-4 http://oreilly.com/catalog/relax (W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema ------------------------------------------------------------------------
Received on Tuesday, 8 August 2006 21:36:30 UTC