Rambling about the Lower Case Semantic Web

Danny Ayers has a rambling blog entry about Microformats (a.k.a. the Lower Case Semantic Web). It's a disjointed read, nevertheless it prompted me to think about this in a bit more detail.

The Technorati Developer's Wiki is a pretty extensive definition of Microformats. Microformats in general is a semantic web like format that uses only existing XML schemas like XHTML. The obvious thought that pops up when one sees this is "Isn't XML designed to be extendable?". That is, why should one restrict oneselves to use only existing XML schemas?

The answer to this can be found by studying how XML evolves in the wild. I wrote a short survey about the current thinking in this space entitled "The Trouble with Evolving XML Schemas". The conclusion was pretty clear: "if you want evolvable schemas, then don't even think of creating new elements". Tim Bray, father of XML, had a similar intuition months ago (sorry, I just can't find the reference URL).

RDF and its encoding has always been the de-facto standard for representing the semantic web. Almost everyone who has looked into RDF has complained about the readability of the XML syntax (one clearer alternative is N3). That was my main stumbling block years ago when I studied it. Bill de hÓra has an good summary of the current state of RDF since I also last looked at it.

Now Tim Bray is proposing yet another simplified syntax (also a nice historical view of RDF and an RDF challenge ). He writes in the spec:

1. The syntax of RDF/XML is sufficiently scrambled and arcane that it is neither human-writeable nor human-readable.
2. The RDF/XML syntax makes heavy use of qnames that is neither intuitive to humans nor conforms particularly well to Web Architecture, which requires that everything significant be identified by URI.
3. People who care about metadata have no trouble thinking in terms of resource/property/value triples.
4. Alternatives like N3 that make the RDF triples evident in syntax suffer in comparison to the XML/RDF syntax because they lack XML's widely-deployed base of software, i18n facilities, and APIs.
5. The notion that you RDF can be mixed into XML transparently enough to be unobtrusive has failed resoundingly in the marketplace.

It's definitely a much simpler version. However, let's go a step further and get rid of the R,PV elments and embed this in the Microformat way. For example, the assertion:

<R r="/resources/rA">
 <PV p="/properties/pA" v="/values/vA" />
 </R>

will instead look like this:

<LI>
  <A rel="P" href="/resources/rA"></A>
    <DL>
     <DT>PV</DT>
            <DD>
                <UL><LI><A href="/properties/pA">/values/vA</A>
                       </LI>
                 </UL>
            </DD>
    </DL>
  
</LI>

I haven't worked out the complete details or cleaned up the syntax, however I hope you get the idea. It's pretty verbose right now. I'll need to dig into some more exampes to see if it can be made much cleaner. The advantage of encoding RPV in lower case style is that you don't need XMDP. The advantage over RPV is that you only need a write CSS to come up with a viewable form. RPV will require one to work with XSLT and Javascript to have something viewable.

I don't know yet what will come of this, nevertheless it looks like #6 in my predictions for 2005 is gaining traction.