- From: Maxime Lefrançois <maxime.lefrancois@inria.fr>
- Date: Wed, 13 Jun 2012 11:04:48 +0200 (CEST)
- To: Yves Savourel <ysavourel@enlaso.com>
- Cc: public-multilingualweb-lt@w3.org
I don't think there is a PROV-notation example that replaces the usage of the ITS idValueRule rule as I understand it. ----- Mail original ----- > De: "Yves Savourel" <ysavourel@enlaso.com> > À: public-multilingualweb-lt@w3.org > Envoyé: Mercredi 13 Juin 2012 10:58:20 > Objet: RE: ISSUE-9: idValue > > Thanks for the pointer Maxime. > > I've read the definition, but I'm afraid I still don't see how > concretely prov:hadOriginalSource would be used in an ITS rule. > That's probably me not understanding how PROV works. > > PROV seems to be about expressing relations, but here we don't have a > relation yet between the original content and the extracted one. > > Could you maybe write my small example with the PROV notation, so we > can picture how that would look like? > > -ys > > -----Original Message----- > From: Maxime Lefrançois [mailto:maxime.lefrancois@inria.fr] > Sent: Wednesday, June 13, 2012 9:26 AM > To: Yves Savourel > Cc: public-multilingualweb-lt@w3.org > Subject: Re: ISSUE 9: idValue > > Hi Yves, > > I might not fully understand all the needs/implicatures of the > idValue, nor how that would work in practice exactly, I suggest you > have a look at the definition of the prov:hadOriginalSource relation > there: > http://www.w3.org/TR/2012/WD-prov-dm-20120503/#term-original-source > > This relation could be used to link the str:String Entity of the > output to the str:String Entity in the original document. > - This is orthogonal to any other data category, > - This don't replace the possibility to apply the best practice that > consists in giving the same local id than in the original document. > > Maxime. > > ----- Mail original ----- > > De: "Yves Savourel" <ysavourel@enlaso.com> > > À: public-multilingualweb-lt@w3.org > > Envoyé: Mercredi 13 Juin 2012 10:01:42 > > Objet: ISSUE 9: idValue > > > > Hi all, > > > > To follow up on the possible idValue feature: > > > > I may not completely understand how PROV works yet, but it seems to > > me > > this would apply to store the id after it has been extracted. > > > > So whether or not it's used in the output, it wouldn't change the > > need > > for an ITS rule for indicating how to acquire the ID values from > > original documents where you have existing IDs. > > > > I also think this should be its own data category rather than an > > attribute in the translateRule (as the current extension proposal) > > because application may need to use the feature without using the > > Translate data category (for example with the Terminology data > > category). > > > > We would have for example: > > > > <?xml version="1.0"?> > > <messages> > > <its:rules version="2.0" > > xmlns:its="http://www.w3.org/2005/11/its"> > > <its:idValueRule selector="//msg" idValue="@num"/> </its:rules> > > <msg num="123">Some text for ID 123</msg> </messages> > > > > The local attribute xml:id would override the global rule, the same > > way xml:lang is the local attribute for the Language Information > > data > > category. > > > > Cheers, > > -yves > > > > > > > > > > >
Received on Wednesday, 13 June 2012 09:05:17 UTC