RDF Blog

A blog covering the Resource Description Framework, in plain English.

Website semantics, choose your weapon

21/10/2008 11:47
posted by chris ward.

Found an intriguing tool from a quick google today, a score matrix for allowing you to choose the right semantic-web technology for your app.

In the line-up is Microformats, eRDF and RDFa ...

Compared for the following features, each technology is scored for your requirements;

  • DRY (Don't Repeat Yourself)
  • HTML4 / XHTML 1.0 validity
  • Custom extensions / Vocabulary mixing
  • Arbitrary resource descriptions
  • Explicit syntactic means for arbitrary resource descriptions
  • Supported by the W3C
  • Follow DCMI guidelines
  • Stable/Uniform syntax specification
  • Predictable RDF mappings
  • Live/Web Clipboard Compatibility
  • Reliable copying, aggregation, and re-publishing of source chunks. (Self-containment)
  • Support for not just plain literals (e.g. typed dates, floats, or markup).
  • Triple bloat prevention (only actively marked-up information leads to triples)
  • Possible integration in namespaced (non-HTML) XML languages.
  • Mainstream Web developers are already adopting it.
  • Tidy-safety (Cleaning up the page will never alter the embedded semantics)
  • Explicit support for blank nodes.
  • Compact syntax, based on existing HTML semantics like the address tag or rel/rev/class attributes.
  • Inclusion of newly evolving publishing patterns (e.g. rel="nofollow").
  • Support for head section metadata such as OpenID or Feed hooks.

With reports of the scoring being biased towards RDFa (well of course, it's extensible stupid!), there is a caveat that it probably isn't the one to go for, for lack of a W3C Recommendation "Spec".

But since last tuesday, that should be an invalid argument now... I'd love to see this document updated to reflect that, Benjamin? :)

Either way, fill your boots ;)

Semantic advertising, the next step?

20/10/2008 15:36
posted by chris ward.

Well, you have contextual advertising (advertising based on the content of a page) and behavioural advertising (based on the end-user's usage of the site)...

The problem with advertising based on context, is that computers are often unable to understand what we mean, and so therefore will never be 100% accurate.

And the problem with behavioural advertising is that it's plain intrusive, disregards a user's privacy, and can be easily thrown by more than one person using the same machine, in any given period of time, xmas shopping on the home family-shared pc, for example.

Maybe with added semantics, the advertising put in place would seem more relevant to the user, more effective.

CURIEs; Compact URIs

20/10/2008 14:19
posted by chris ward.

Sometimes RDF documents can scale to be rather large and unweildy, this is partly caused by the placement of long URIs for the RDF nodes.

Fortunately this problem can be overcome by using CURIEs in place of the URIs.

Short for Compact URI, CURIEs abbreviate URIs!

Found a wicked example and explanation over on wikipedia.

As explained in the article, and also the W3C's working draft for CURIE Syntax, it's also handy to note that to avoid disambiguation between URIs and CURIEs, such collisions and confusion can be avoided by wrapping CURIEs in square brackets;

<html xmlns:ex="http://www.example.org/">
    <head>...</head>
    <body>
        <p rel="foaf:homePage" about="[ex:home.html]">home</p>
    </body>
</html>

URIs; URLs and URNs

20/10/2008 14:01
posted by chris ward.

An identifier for an object in a given domain or namespace, can be presented as a URL or a URN.

A typical representation of a URI can be found in your every-day website address.

   http://rdfblog.com/blog/URIs-URLs-URNs

Originally URLs were designed as an informal concept for identifying objects based on location within a domain, but also proved to become quite ideal for linking resources.

URNs can identify objects based on name, for example;

	urn:isbn:n-nn-nnnnnn-n

"http" and "urn" are both URI schemes, in a list of 84 that are publically available, around 30 are registered.

However, this list has now been retired, but fortunately I've discovered a wiki that covers URI registration.

Many of the references to Objects, Subjects and Predicates in RDF use URIs to represent their nodes, so hopefully this post helps to clear the air.

Various formats for presenting RDF statements

20/10/2008 12:29
posted by chris ward.

Because RDF is a concept and not a syntax, there are many ways to present RDF statements to a user or user-agent;

  • RDF Graph
  • N3
    • N-Triples
      • TURTLE
  • XML-RDF
  • XHTML+RDFa
  • eRDF

I will try to summarise the above in a series of posts

RDFa becomes a W3C Recommendation

20/10/2008 11:52
posted by chris ward.

Officially now a 'web-standard', RDFa is the implementation for writing RDF statements inline to a XHTML document using 9 specific attributes.

4 of these are supported in XHTML 1.0

  • rel
  • rev
  • href
  • src

The other five are specific to XHTML+RDFa 1.0, so your document will have to conform to the DTD.

  • about
  • property
  • resource
  • datatype
  • typeof

XHTML+RDFa-1.0-DTD

20/10/2008 11:35
posted by chris ward. 

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML+RDFa 1.0//EN"
 "http://www.w3.org/MarkUp/DTD/xhtml-rdfa-1.dtd">

Search Monkey overview

17/10/2008 10:27
posted by chris ward.

purpose of this blog

01/10/2008 15:31
posted by chris ward.

Welcome!

This is the first blog post of many regarding RDF, the Resource Description Framework.

The blog will cover the different respresentations of RDF statements, N3, RDF-XML, RDFa, eRDF, etc etc.