[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office-metadata] Fw: News Release: W3C Completes Bridge Between HTML/Microformats and Semantic Web with GRDDL
I know the GRDDL group pretty well, because Harry Halpin who co-wrote the spec is here at Duke. I agree with Rob that GRDDL is a different take on binding RDF to XML, and is neither good nor bad news from the perspective of ODF- metadata. My own view of GRDDL (and I've told Harry this) is that it's a fantastic idea for bridging traditional concrete XML vocabs and the RDF world, but if you have good support for RDF built into an application from the beginning, it's not a critical technoloty. It's a transitional technology, in my view. If my app has good built-in RDF support, instead of creating a microformat, I'd just create a micro-ontology in RDF/OWL, and reference that. (Some GRDDL people come from the microformats world rather than the ontology world, and they think that microformats are more likely to get permanent traction than micro-ontologies in the real world. I'm not convinced.) Incidentally, just to give another take on this: one of the ODF metadata options we are offering is "inline" metadata, which is like RDFa. And I think Harry agrees that RDFa is, in a sense, a GRDDL standard because it (a) establishes a special XML vocab that is (b) intended to be extracted into RDF (c) using a publicly-available transform--the same idea behind GRDDL. It's just that in RDFa (as in ODF-inline) the transform that is used to extract the RDF is not domain-specific. Which reminds me, at some point we will need to write (and publicly release) an XSL that OO can use internally (and others who access content.xml files outside of OO can use) to extract and transform ODF- inline metadata markup to RDF. This should be a pretty simple XSL to write. John On Sep 12, 2007, at 4:51 PM, Svante Schubert wrote: > robert_weir@us.ibm.com wrote: >> >> I guess I was thinking along the lines of spreadsheet data. Users >> commonly put in data tables, a few columns of data with labels as >> headers. These may start in row 1, or anywhere. If another >> application wants to operate on that data in the document, they >> really don't know where to start. What range do they look for? >> What are the data types of the columns? It isn't really self- >> describing because the column headers are really not headers in >> any formal sense. They are just cells containing labels. It >> looks nice to the eye, but it really isn't formally structured. >> >> >> So what would you do if you wanted to layer some structure on to >> the data? I guess you could annotate with RDF metadata and say >> things about the range and the headers. But another way might be >> to define via GRDDL a way to extract RDF from the spreadsheet. So >> not really metadata, but RDF to describe the actually data of the >> sheet. This could be powerful. > The structure you described in your example should be described by > RDF, from my understanding GRDDL does not add any structure. GRDDL > only shows a way how to extract the RDF graph from a XML file by > providing an algorithm (usually XSLT stylesheet). > > The RDF graph of an ODF document should be the same in the end with > or without GRDDL. > Or did I misunderstand your scenario? >> >> But I agree that this doesn't necessarily intersect with our >> metadata work. Both use RDF, but for different purposes. >> >> -Rob > Bests, > Svante >> >> ___________________________ >> >> Rob Weir >> Software Architect >> Workplace, Portal and Collaboration Software >> IBM Software Group >> >> email: robert_weir@us.ibm.com >> phone: 1-978-399-7122 >> blog: http://www.robweir.com/blog/ >> >> Svante.Schubert@Sun.COM wrote on 09/12/2007 03:18:11 PM: >> >> > Bruce D'Arcus wrote: >> > > robert_weir@us.ibm.com wrote: >> > > >> > >> Does this mean anything for us? > > >> > > It might. It would allow, for example, someone to include some >> custom >> > > XML in an ODF package, but include the GRDDL logic to convert >> it to RDF. >> > > >> > > I'm not really sure of the practical use case for that, but I >> imagine >> > > someone may find one. >> > > >> > GRDDL is a framework to extract RDF from plain XML by providing an >> > algorithm (usually a reference to a XSLT stylesheet). >> > We in ODF 1.2 offer more than a framework for plain XML, we offer a >> > framework for package formats using XML files. >> > >> > Instead of a GRDDL algorithm we rely on a fixed algorithm by in >> general >> > annotating the relevant XML resources (element nodes) by xml:id >> > attributes and using one metadata manifest to relate xml:ids >> with IRIs >> > and to describe related RDF files of the package. >> > >> > Aside of this package metadata mechanism we define typical ODF >> resources >> > (e.g. odf:ContentFile) and in content metadata of an ODF file (e.g. >> > m:about, m:property). >> > Making this ODF in content metadata extractable using GRDDL and >> XSLT >> > seems possible, but would require the interpretation of the >> manifest and >> > providing the XSLT stylesheet. >> > >> > Not sure if it is worth the effort, as I doubt that a common >> GRDDL agent >> > would even unzip an ODF document at all. >> > I would further assume GRDDL would drop some of our nice >> features as RDF >> > named graphs like we currently have for each RDF file. >> > >> > Rather than looking on agents, I am more focused on the Office >> Extension >> > developer, which might rely on some XSLT library as RDF TWIG when >> > working with ODF metadata using XSLT. ;-) >> > >> > regards, >> > Svante (on vacation but could not resist) >> > >> > [1] http://rdftwig.sourceforge.net/paper/index.html >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]