OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   Re: [xml-dev] Why RDF is hard

[ Lists Home | Date Index | Thread Index ]

Simon St.Laurent wrote:

> jonathan@openhealth.org (Jonathan Borden) writes:
> >Sorry, I was refering to RDF itself, as opposed to RDF/XML. the much
> >maligned XML serialization syntax.
> So was I.  The serialization syntax merely makes a bad situation worse.

Nah, the concept of triples and directed labelled graphs (DLGs) is really
pretty basic -- but it's not one that is useful to all problem domains by
any stretch. Either it is useful to you or not -- simple as that. Now DLGs
certainly aren't the most basic type of math but the idea of functions which
take one argument and have a value (which essentially is what triples
represent) is pretty basic.
> N3 makes a bad situation clear.  That clarity doesn't help those of us
> who don't think in triples and don't have the patience to string triples
> together into structures.
> For those who think in triples, RDF is a delight.  I just wish those
> people would recognize that a lot of people cannot and will not be
> joining them in that delight.

That's sort of like saying that you don't think in terms of algebra -- which
is probably true. The point is, that unless the problem you are trying to
solve is a good fit for a graph type of analysis, then there is no reason to
"think in terms of triples". I don't think in terms of triples either, any
more than I think in terms of characters, or think in terms of lists, or
arrays etc.

Generally I do think in terms of classifications, which is why I find
ontologies useful, but there is no necessary need for ontologies to be based
on triples -- indeed there are many many on the WebOnt WG who would prefer
to think in terms of lists ... the old Lispers in fact ... ANSI KIF is
another knowledge representation language built on s-expressions in fact.

So for me, triples are but a means to an end, sort of like Java bytecode --
something that the designers and implementers of Java need to be aware of,
but not something that I personally think about on a daily basis.

In any case there is indeed an "impedence mismatch" between triple based
DLGs and tree based XML and this the root of the essential difficulty with
XML and RDF.

On the other hand you should see that folks who deal with classifications,
including classifications of documents and mappings between various systems
of classifications, find RDF type systems helpful.  For example John Cowan's
day job -- classifying healthcare articles and my work classifying medical
records and diseases.



News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS