A Comparative Clarification: Microformats vs. RDF
James Simmons posted a blog entry on the Semantic Focus blog that outlines his thoughts on the differentiation between Microformats and RDF as they relate to the semantic web, and Johannes la Poutré replied on the Squio blog replied with his own interpretation. While they both agree that RDF and Microformats are very different, they have a very different take on how that impacts their respective relevance to the semantic web. James started by pointing out that Microformats intentionally keep out of scope some core aspects of RDF:
Many people have contended that Microformats are a possible replacement for RDF, however Microformats were not designed to cover the same scope as RDF was. While both Microformats and RDF make it possible to store data about data, they simply do not work to solve the same set of problems. ... Directly on the Microformats about page you are told what Microformats are and are not.
What Microformats were not intended to be:
- A new language
- Infinitely extensible and open-ended
- An attempt to get everyone to change their behavior and rewrite their tools
- A whole new approach that throws away what already works today
- A panacea for all taxonomies, ontologies, and other such abstractions
- Defining the whole world, or even just boiling the ocean
In contrast, RDF has a much more ambitious scope, and in James' opinion some of those things are critical to the semantic web:
What RDF allows (and Microformats lacks):
- Resources are represented as URIs, allowing you to access metadata remotely
- Infinitely extensible and open-ended design
- A powerful Ontology language (OWL) that is built upon it
- The ability to utilize, share, and extend any number of vocabularies
- No reliance on pre-defined "formats" (i.e. not limited by the types of data that can be encoded)
As you can see there are a few things we can do with RDF that cannot be done with Microformats. The Semantic Web relies on the things I've listed above. These are the clear-cut reasons why Microformats will not be part of the W3C's Semantic Web vision.
Taking the compare/contrast analysis a bit further, James also described some things that Microformats have that RDF lacks, and describes how he sees Microformats contributing to building the semantic web.
I'd like to say that Microformats do have a number of qualities that RDF (although not necessarily all serializations) does not accommodate for, at least not in the same way:
- Designed for humans first, machines second
- Modularity / embeddability
- Enables and encourages decentralized development, content, services
- A design principles for formats
- Adapted to current behaviors and usage patterns
- Highly correlated with semantic XHTML
I've stated before that I believe Microformats will help bring about the Semantic Web by introducing "metadata sprinkling" (the act of including metadata in otherwise "normal" data) to more people. They allow for simple metadata embeddability and do not affect how an XHTML document validates. This is the kind of approach that will help normal users come closer to understanding the Semantic Web vision.
Joe took issue with the characterization that the humble scope of microformats lessens the impact or role that microformats will have in defining the semantic web. But he started by agreeing with James' differentiation between the two concepts:
Indeed, microformats are not an alternative for RDF, not even a “poor man’s version”. But that was not a design goal at any time. What’s more: microformats are no first class semantic web objects in any way either. Rather, they are the simplest imaginable solution for semantically correct markup, limited to the most common data formats out there.
To rephrase the microformats charter, they want to be the common man’s solution, aimed at the well intending webmaster crowd. As such, microformats can be hugely successful (analogue to the “html as tag soup” success story). Fine.
He then characterizes the RDF crowd as lofty and out of touch, and puts forward his own view of the future, where microformats will play a larger role that RDF in defining the semantic web because of it's attainability.
Bottom line: the semantic web has been lacking real world content for too long (not withstanding DBPedia and Freebase and such) and real world applications for the common man. Microformats can and will have a place in advocacy for this large target audience, people who grasp html and basic data constructs, but who are not interested in graph theory.
This audience will only jump on the bandwagon if they can instantly understand the intent from view source inspection. Compare the success of RSS 2.0 over the semantically superior (but more complex, RDF based) RSS 1.0 version.
In the end it will just not matter, most content will be “good enough” to be useful for the semweb (through GRDDL transformations and screen scraping), just like today’s html is good enough to be rendered, in some way, in our web browsers.