Idea #19: Take a look at MicroFormats

One of the more interesting technologies that I’ve noticed emerging in the past year or two is the concept of Microformats.  Essentially what microformats provide is a standard format for embedding XML data in textual content.  What this allows us to do is embed actual data in content and then automate data extraction.

There are already a ton of companies implementing this technology–you can use it today.  A lot of semantic Web people don’t like it because it’s far more unstructured than technology like RDF.  However, you can’t argue with the fact that it’s being used.  And the main reason, in my opinion, is that it gives people a concrete method to use them, unlike RDF, whose definition is very hazy when you try to wrap your brain around it.  There are a ton of standard microformat schemas that you can use and they’ll start working immediately.

For an example of what microformats can do, you may want to check out the Microformat viewer I posted a while back.

Part of the 60 Ideas in 60 Days series.  Click here for the rest of the ideas.

Share and Enjoy:
  • Print
  • Digg
  • Facebook
  • Google Bookmarks
  • HackerNews
  • Reddit
  • http://dannyayers.com Danny

    Hmm, "hazy" wouldn't be the first word I'd use for RDF (it does have mathematically defined semantics). But I think I know what you mean – it can seem like there's too much flexibility, hard to pin down at times.

    Fortunately there's no need to choose between microformats and RDF. Microformats can be mapped to RDF vocabulary, and there's a set mechanisms by the name of GRDDL to allow automatic extraction, without the person publishing the HTML having to go anywhere near RDF. (Some Semantic Web folks may be critical of microformats, but others are very enthusiastic. And vice versa :-)

    It's be possible to build viewers and the like directly on top of microformat data. But what RDF brings to the party is the ability to merge and query data from multiple sources, from multiple microformats, without having to write tons of format-specific code.

    GRDDL is still a work-in-progress, but there will almost certainly be demos (at least) involving microformats used in this way in the near future.

    'Course it's also possible to go the other way, e.g. publish the results of SPARQL queries as microformat-enhanced HTML. But the big gain (and fun) comes from using RDF as a common data model behind microformats (or any other data/formats).

    See also: micromodels

  • http://www.jasonkolb.com Jason Kolb

    Thanks for the reply Danny, I skimmed the GRDDL spec and it does indeed look promising. I'll give it a more thorough run-thru over the next week. Any thoughts on how to do two-way data with RDF (inputting data as well as reading and searching data)? I haven't been able to find any work being done in that area and I think it's pretty critical before any "real" apps are written using RDF. Or, in your mind, is RDF more for exploring and discovering application output rather than being involved in the data conversation that puts data into the system?