[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office-metadata] Linking in a vocabulary
Elias, Elias Torres wrote: >Patrick Durusau <patrick@durusau.net> wrote on 05/04/2007 10:42:22 AM: > > > >>Elias, >> >>Snipping to the request for more information: >> >>Elias Torres wrote: >><snip> >> >> >> >>>>Second question: How do I say that a term that is defined by more >>>>specific metadata, through inline metadata association, should use that >>>>triple and not the more general one that would apply to the document as >>>>a whole? (Or is that something that we need to say in the proposal? >>>> >>>> >That > > >>>>inline metadata trumps vocabulary metdata applied to the document as a >>>>whole? Well, more formally than that but you get the idea.) >>>> >>>> >>>> >>>> >>>I don't think anything trumps anything. At the end, all we are doing is >>>generating triples from all of the places in our package. We can have >>> >>> >some > > >>>provenance of where each triple came from, but they can all co-exist in >>> >>> >a > > >>>single graph. It really depends on your use, which trumps which. No >>>normative trumping in our spec. >>> >>>However, I'd need more help understanding your scenario because it's a >>> >>> >bit > > >>>too high level for me at this point. Could you try first to give us some >>>sample data you are trying to model and then we can figure out how to >>> >>> >use > > >>>the spec to encode it? >>> >>> >>> >>> >>> >>Well, assume that I have a Bible vocabulary for all the proper names in >>an English translation of the Bible. >> >>That means that I have entries for (not complete): >> >>Joseph - advisor to the Pharoah >> >>Joseph - a musician in the service of David >> >>Joseph - husband of Mary >> >>Joseph - father of Jesus (as seen by his contemporaries) >> >>I could, of course, represent those subjects with a set of RDF >>statements that delimit the verses where they appear. >> >> > >Right. Something like this, right? > ><text:p>Text from the Bible....more text<text:span >m:about="urn:joseph-son-of-jacob" m:property="foaf:name">Joseph</text:span> >more ...</text:p> > > > Yes, but note that requires marking *every* string "Joseph" with the <text:span> element. >>In other words, a set of RDF statements about Joseph -advisor to the >>Pharoah, plus Gen. 37:1-47:27, so that term will only be interpreted as >>that Joseph within that verse range. >> >>But the problem of having the same string that represents different >>subjects occurs in texts where that "easy" solution isn't possible. >> >>In other words, what if I have two separate vocabularies, one for Cato >>(the elder) and Cato (the younger). In the context of a scholarly >>article about Cato (the elder), if I don't do anything, that is the >>triple(s) that should apply to any mention of "Cato." But, from time to >>time, I want to mention Cato (the younger) and that should draw metadata >>from a second vocabulary, perhaps the OCD (Oxford Classical Dictionary), >>which Bruce has kindly encoded in RDF. ;-) >> >> > >This you mean in the manifest file.... > ><odf:ContextFile rdf:about="urn:doc-id" odf:path="context.xml"> > <patrick:refersToVocabulary rdf:resource="urn:first-vocabulary"/> ></odf:ContextFile> > >So an application can first go to the ContextFile and see which is the main >vocabulary to search when looking for people's names. But then you confuse >me.. what about a second vocabulary. I'm assuming you are willing to >specify something at least in the content.xml that points to another >vocabulary. > > > No, an application can look at the manifest first to discover the vocabulary that applies to *all* strings in the text, except those that have explicit metadata pointers associated with them. That is I would have metadata associated with strings in content *without* having to mark those strings without having to make a reference to the metadata from an element surrounding the string or by binding a bookmake using xml:id. Ah, no, I wasn't going to specify something in content.xml that points to the 'default' vocabulary. I only have to point out of context.xml when I want a particular part of some metadata file to be associated with a particular place in the text. ><text:p xml:id="foo">Text...more text<text:span>Cato</text:span> more >...</text:p> > >Then you would add something in RDF/XML that says xml:id="foo" should use a >second vocabulary. Is that what you want? > >It would have been nicer to just put the triple right in the doc with the >meta field. > ><text:meta about="" property="p:overrideVocabulary" >resource="urn:second-vocabulary"> >put your text over here. ></text:meta> > >But I think we agreed for now to just use RDF/XML to encode that >information. > >Does this help at all? > > > Well, good examples of pointing to metadata but my question is about having metadata associated with content without having to do any explicit pointing. That is I have a default vocabulary that is applied to every matching string. I don't think I am doing a good job explaining this. Bruce, are you around? I will try to work on it over the weekend so I can have an example that is formally cast into the right form. Maybe that will help. Hope you are looking forward to a great weekend! Patrick >-Elias > > > >>When I said "trump" what I meant was that the more specific metadata, >>that which is associated inline, is used in preference to the general >>metadata, which I have associated with the entire file. >> >>Think of it as being the same as an inherited attribute value where >>inheritance is blocked by the specification of a specific attribute >>value. (Close as I can come to a markup example.) >> >>The problem is that I don't know if we should say that in the proposal >>or if not, how to say that in the metadata. >> >>There will be some documents that only use one vocabulary with no >>conflicts but I suspect that is going to be the exception rather than >>the rule. >> >>John Madden can confirm if that is going to be the case for medical >>documents, but I suspect it will be true. >> >> >> >>>>Hope everyone is looking forward to a great weekend! >>>> >>>> >>>> >>>> >>>Lots of home projects. Yeah! >>> >>> >>> >>> >>> >>Hopefully not! I have an ISO draft I have been promising for weeks now >>that is top of the weekend stack! ;-) >> >>Hope you are having a great day! >> >>Patrick >> >> >> >> >>>>Patrick >>>> >>>>-- >>>>Patrick Durusau >>>>Patrick@Durusau.net >>>>Chair, V1 - Text Processing: Office and Publishing Systems Interface >>>>Co-Editor, ISO 13250, Topic Maps -- Reference Model >>>>Member, Text Encoding Initiative Board of Directors, 2003-2005 >>>> >>>>Topic Maps: Human, not artificial, intelligence at work! >>>> >>>> >>>> >>>> >>>> >>>> >>> >>> >>> >>> >>> >>> >>-- >>Patrick Durusau >>Patrick@Durusau.net >>Chair, V1 - Text Processing: Office and Publishing Systems Interface >>Co-Editor, ISO 13250, Topic Maps -- Reference Model >>Member, Text Encoding Initiative Board of Directors, 2003-2005 >> >>Topic Maps: Human, not artificial, intelligence at work! >> >> >> >> > > > > > > -- Patrick Durusau Patrick@Durusau.net Chair, V1 - Text Processing: Office and Publishing Systems Interface Co-Editor, ISO 13250, Topic Maps -- Reference Model Member, Text Encoding Initiative Board of Directors, 2003-2005 Topic Maps: Human, not artificial, intelligence at work!
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]