RDF (ThoughtStorms)

Format which aims to break all meta-data down into a simple tuple (subject, predicate, object) where both subject and object can be URIs.

TimBray : http://www.tbray.org/ongoing/When/200x/2003/05/21/RDFNet

: I notice two years later it doesn't look like the challenge has been won.


BillDeHora has trouble with the markup :

New thinking : http://www.ilrt.bris.ac.uk/discovery/2003/11/new-syntaxes-rdf/paper.html



Now I'm trying to get into this for OptimaesCurrencyDescription.


In a sense it's pulverising data into a microscopic set of predicate claims (See also BangTheRocksTogether)


Here's my complaint against RDF-XML (and RDF in general)

If RDF-XML is just an XML serialization of RDF triples why doesn't it look like this? :

URI

URI

URI

Seems the reason it doesn't is some kind of PrematureOptimization thing.

doh! It's "R" as linked from this page (above) several months ago. And I forgot!



a lot of rdf related projects and ideas I don't understand : BillSeitz:z2002-06-06-a

strikes me that this stuff is for ArchitectureAstronauts


It's the syntax. But maybe the real problem is the URL as unique ID. There's a reason for this, of course. But if you were starting out to create a way of representing UIDs, would you start with something that looks like a URL? That always had to start http://? http:// is an access protocol? So what would happen if I used ftp:// with the same location? Does this refer to the same thing? Wouldn't a URI have been better to follow the Java package convention : com.nooranch.application.submodule etc?

The more I think about this, the more pointless the similarity of URIs to URLs is. The URL is a way of "getting" or moving something through the web. The URI is just a technique for uniquely labelling it in a fixed place. Making URIs look like URLs is tantamount to deciding that houses should be backwardly compatible with cars. And so have doors the same shape. Fixed things don't have to look like moving things.


Contrast :

See also :