[rdfweb-dev] Time's running out?

Danny Ayers danny666 at virgilio.it
Fri Aug 8 10:59:42 UTC 2003


> And I'm still extremely confused about how anyone is supposed to consume
> this stuff. It looks to me suspiciously like write-only data.

Get a ready-made parser and create a list of (triple) statements, or a graph
structure. Do stuff with that.

> And even with relatively simple concepts, it has that perl quality of
> "There is always more than one way of coding". To take a single example,
> I was amazed that something as simple as encoding a Person's gender
> should have produced more than a single proposal. I can understand
> arguing about the name of the tag, but why was there any argument about
> structure?

There are (at least) two sides to this - one is the RDF model, and the
structure there; the other is the XML serialization. From a data modelling
point of view there are two or three different alternatives, how the
information fits in the RDF model as a (sub)graph/list of statements. The
most obvious choice being between a literal "female" or it could a term
defined by a resource : http://wordnet.org/stuff#female, but then there is
also the possibility mentioned of creating foaf:Female and foaf:Male classes
as subclasses of foaf:Person. But once a choice has been made, it makes
sense to stick to it ;-)

The same structure in the RDF model can have different representations in
its XML representation. Put these together and there are quite a few
alternatives, but it should be remembered that structural difference in the
model and the difference in the syntax are different issues.

> I think we have to take a 10,000 ft view once in a while and think about
> adoption and the lifecycle of standards. Every namespace and every tag
> within that namespace is a "standard" that has to fight for survival in
> the ecology of developer and application mindshare. If you control both
> ends (author and consumer) this is just not an issue. And if you want to
> expose it globally but don't care if anybody reads it, again, no issue.
> But if you want to get widespread adoption, then the standard has got to
> be easy to produce, easy to understand, easy to consume with code. These
> are just pre-conditions. You've still got to do all the evangelism and
> get lucky.
>
> So is FOAF and by implication RDF, "Write-only" data? Well from where
> I'm standing, damn right. Using the tools available in my chosen
> environment (PHP, not exactly off mainstream) It's dead easy to create.
> It's hard to read.

How so? As above - get a parser off the shelf and use it to create a list of
(triple) statements, or a graph structure. Do stuff with those/that.

And lots of it I can't read reliably at all or
> interpret if I can read it.

The reliability is essentially the same issue as any XML data. The
interpretation is down to what you want to do with it. Do you have some
examples?

And it's taken me 8 months off and on to get
> to the point where I can mostly understand the issues.
>
> This is all causing me extreme frustration and sleepless nights ;-) I
> think I can vaguely see a huge playground somewhere in the future. But I
> have no idea how to get there. And I have this nagging doubt that we
> can't get there from here.

If you described the specific problems you're experiencing, I think it
likely someone here has encountered them before and can help.

Cheers,
Danny.




More information about the foaf-dev mailing list