You have your peanut butter in my chocolate!

Recovered from the Wayback machine.

Jon Udell has been exploring the concept of mixing, in his words, RDF-isms with RSS 2.0, which is a non-RDF, single use XML vocabulary.

First, important note – when RDF people talk about RSS, they usually mean RSS 1.0, which is an RDF-enabled vocabulary.

Second important note – RDF puts certain constraints and requirements on an XML document for it to be valid RDF/XML, and to be usable within RDF applications and APIs.

The question began with namespaces, and Dan Brickley pointing out what is one of the major strengths of RDF – if a vocabulary is RDF compliant, then it’s namespace would work within other RDF-compliant vocabularies. As Dan wrote in comments in the original thread:

In the RSS1 design (via the love-it-or-loathe-it RDF approach) we had a more loosly coupled, de-centralised design: a namespace worked with RSS1 if it worked with RDF. If someone created an RDF vocab for Jobs, then it worked with RSS1. If someone else creates an RDF vocab for locations (to talk in more detail about where the jobs are), then that too just worked. Same goes for a skills vocab (there’ll likely be several). Or person-descriptions (not just FOAF, vCard but the several others people have created to qualify those).

Jon took this tidbit and stretched it – waaaay out of shape – into a discussion about using RDF in conjunction with RSS 2.0. He comes up with a sample feed. It validates with the RSS validator. Fine and good – but it’s not going to validate as RDF, which provide those aforementioned rules and criteria and constraints that keep any old XML from being used as RDF and violating the RDF model that forms the basis of the RDF/XML.

There is a mathematical model we all use that says when we add 2 plus 2, we get 4. In addition, within this model a value of ‘3′ is greater than a value of ‘2′. You can create your own mathematical model, and in it, 2 plus 2 could equal 5, or 3, or 6 if you want, and a value of ‘2′ is greater than ‘3′ – but it won’t work with the existing model.

You could take your model and document it nicely and say, “Please use it. It’s better and more simple”, and you might be able to get people to use it – but then they would never be on time for trains, and they would never have correct change, and they would most likely be in trouble with the income tax folks.

The point is that the numbers, the operators, and even the syntax that we use within our mathematical system isn’t what’s important – it’s the model, not the pictures that counts.

Jon today states:

Actually, I’m not saying that I want to put RDF into RSS. I’m trying to ask and answer two questions: 1) Is it feasible? and 2) What benefits would it confer?

Jon, in a nutshell: 1) no , 2) many, but would require that you abandon the single-use RSS 2.0 architecture in favor of a more universally defined architecture, which is the aforementioned RDF/XML.

Lots of new RDF/XML vocabularies coming down the road, Jon, and not just the new RSSJobs which is this week’s hot spot. (Is this RSS 1.0 or RSS 2.0 – can’t find info on this.) Many excellent RDF/XML vocabularies, all of which will not work with RSS 2.0 because RSS 2.0 is a single-use, single-purpose simple syndication feed that doesn’t care whether it mixes with other vocabularies. That’s cool. That’s your choice.


Did find a comment thread where Dan talks about mixing RDF namespaces in with regular XML.

Dan’s a patient man. He wants to allow the XML freeforall world to benefit from RDF/XML. I admire him for this. I disagree with even attempting to do this.

What’s the point of the model, the rigor, the challenges associated with RDF/XML, if we’re going to say, “Oh, well, we don’t want to force you to use our model. Don’t worry your pretty little heads about it – just go ahead and just take what you want, we don’t mind.”

It’s the same as saying use your mathematical model, we’ll learn it so that we can give you the proper change when you buy that beer you so obviously need for doing something so daft.

You want to combine plain vanilla XML and RDF/XML? Fine. Use XSLT. Or a sledgehammer. Or drugs.

Print Friendly, PDF & Email