- From: Dan Brickley <danbri@danbri.org>
- Date: Sun, 14 Jan 2007 22:52:18 +0000
- To: Danny Ayers <danny.ayers@gmail.com>
- Cc: Semantic web list <semantic-web@w3.org>, Brian Suda <brian.suda@gmail.com>
Danny Ayers wrote: > I've just added a page of this name [1] to the ESW Wiki, would > appreciate any pointers. While ConverterToRdf [2] lists tools that are > of immediate interest to people working with RDF, conversion the other > way may prove equally important to the web. > > A usage example - a social networking service might offer data import > from vCard format documents. This could be generated from the data in > a FOAF profile via an online service. (Anyone doing that yet? It's > available from microformat formatted data, e.g. [3]). > > Like ConvertersToRdf, there are various different > mechanisms/strategies available. I've listed the following so far: > > * Direct serialisation (Raptor?) > * XSLT over RDF/XML > * SPARQL (one or TwoPhaseSparql) > * programmatic > > The Semantic Web needn't rely on every system understanding RDF (in > fact I'd argue that only a tiny percentage of support is needed to get > major gains in net utility). But RDF/XML (and Turtle) serialisations > and SPARQL endpoints still demand quite a lot of the system at the > other end of the wire. Online extraction/conversion facilities could > ease this burden. Without outputs compatible with existing > applications, there's a danger of Semantic Web applications looking > like black holes. > +1 ... this is a critically important point for the Sem Web community. We very easily make the mistake of seeing "legacy" formats as data sources for scraping, nothing more. In fact they are text-based interfaces to a wealthy of existing, stable, user tested, applications which are perfectly capable of consuming "RDF" data, when reformatted appropriately. I made this point in some slides a couple years ago: http://web.archive.org/web/20050507060219/http://rdf4food.asemantics.org/DanBriSlides - Proprietary/legacy/pre-RDF formats are good things! don't look down on them :) *we often slip into thinking that exif/ldif/icalendar/vcard/etc are useful only as sources of data *in fact, they are simple (a few 'print()' statements...) gateways that allow RDF data to be used in existing applications *deployment model: data flows into RDF from various sources, is merged, integrated, etc., queried, and then flows out again *...eg. generating icalendar files from exif files, via rdf. or running an ldap server from crawled foaf, or a z39.50 server from crawled rss. cheers, Dan > > [1] http://esw.w3.org/topic/ConverterFromRdf > [2] http://esw.w3.org/topic/ConverterToRdf > [3] http://suda.co.uk/projects/X2V/ > > -- > > http://dannyayers.com
Received on Sunday, 14 January 2007 22:52:34 UTC