What eRDF can learn from Microformats

26th February 2007

- 5 comments

Quite a bit of time at BarCamp was spent thinking, talking and in running skirmishes about the Semantic Web. Or the semantic web depending which side you’re on.

Now I’m a big fan and perpetual user of Microformats. They make sense, are simple to add everywhere (by stealth if needs be) and the potential is pretty interesting to boot. They are designed in the open, allowing everyone to participate and have a strong emphasis on solving every day problems in the real world.

Contrast that with RDF and the Semantic Web. Most of the work here has been going on for years in academic institutions and W3C working groups. It’s all rather grandiose. Everything I’d seen previously horribly broke HTML. But someone seems to have seen the light. Tom Morris mentioned in one of his talks eRDF, which is a method for embedding RDF in valid HTML.

After one of the talks I got chatting with Tom Morris and Tom Croucher, both card carrying RDF folk, and Niqui Merret another interested bystander. Toms argument basically comes down to Microformats solve the 80% problem, RDF trys to solve the other 20%. I sort of think the Microformats group would probably agree with that. I also like the idea of creating your own formats off the cuff. I’ve seen enough to at least have a further look see and hopefully try out a few experiements myself.

Ben Ward had a few things to say on the subject which are worth a good read. I agree with the majority of his arguments but coudn’t resist commenting on at least one:

If you’re working with RDF, show me something new. Don’t just rehash vcard and icalendar into your own format; I’m not interested

I disagree. One of the best thinks about Microformats, that draw people (including me) in from the get go, is the unquestionable ease with which you can add hcard or xfn and get on with your day. Weeks or months later, after a few feel good projects you’ll be keeping an eye on the wiki when you come across something in need of structure. And even then you find it’s a composite of something you’ve seen before.

Right now I’d love to find that. A simple example of how to mark up a few very simple examples of eRDF. Hey, if that means contact details and a calendar event then so be it. If all that’s wrapped up in a lovely shiny green coat of pain with an icon logo so much the better. Until I can experiment with simple I cant do anything else, never mind anything new. So that’s two posts for the future sorted then.

Tom has set up a wiki over at GetSemantic to start to collate information together in one place which from my initial research seems like a good idea. If and when I get enough together (and the OpenId authentication is up and running) I’ll make sure to add everything there. If I get carried away and find interesting voices on the blogosphere I might event get a planet up and running. Or I might find way too much XML and leave it at that.

Comments

  1. Hey Gareth,

    good to see you were interested in some of the Semantic Web stuff going on at barcamp. It’s also interesting to see that as one of the points you picked out from Ben’s post.

    I really feel like the main difference between Microformats and eRDF is stuff you get for free. Hopefully the ease of use should be the same (if it isn’t tell us and we should look closer at eRDF), but eRDF should also bring Semantic Web inferencing.

    GetSemantic is going to be a great I think, and a large part of that is going to be finding attractive semantic web examples that are slick and easy to demo. Watch this space!

    Tom Hughes-Croucher - 26th February 2007

  2. [...] Following on from my previous post on eRDF I’ve started to play around with it. For anyone bored enought to have read the source of this site today you’ll have seen a couple of behind the scenes changes – specifically I’ve added a dash of FOAF. [...]

    Morethanseven > Microformats and eRDF sitting in a tree - 3rd March 2007

  3. I think the 80/20 comparison between microformats and RDF formats vastly understates the usefullness of RDF. eRDF can describe as much of the data on your web page as you want; microformats can only describe the most commonly published types. And while microformats makes a virtue of focusing only on the most poular data formats, it ignores the hugely significant long tail of unmet data description needs.

    Keith Alexander - 18th March 2007

  4. I think actually what you describe, that eRDF is great for the long tail, is in complete agreement with Toms original 80/20 comparison. Ben Ward is fond of saying that the limitations of Microformats are designed in, which I think is great. RDF is designed to be useful for every personally concievable use, which makes it more unwieldy that Microformats but also for those times when a Microformat is simply not viable.

    gareth - 18th March 2007

  5. Hi Gareth, if I can try to be a little clearer about what I mean:

    I think the 80/20 comparison might make it seem like microformats are useful to you, the web developer for 80% of your time spent marking things up, while eRDF can only be used 20% of the time (or if you could use it more often, you wouldn’t want to, because you’d be better using microformats).

    Whereas, while microformats might aim to cover most of the most popular types of data structure, you might actually spend most of your time working in the 20% zone.

    I’m not an old-time RDFer with an axe to grind or anything – I used to be really enthusiastic about microformats. But since using eRDF for a bit, I don’t really see the point in microformats anymore. eRDF lets me say anything about anything I want, and in a consistent way (which is just as important). With microformats, I have to learn each new format that comes out, and I don’t get any interoperability at all until the format’s been through The Process, and had some tools written that can parse it.

    YMMV I suppose, but for those reasons at least, I find that it’s actually microformats that’s unwieldy and complex, rather than eRDF.

    So I’m finding it hard to see the benefit, and I’m feeling increasingly skeptical every time I hear a flaw with the microformats approach cited back as a design feature.

    There’s always the tools argument I suppose, but even there, while microformats seems to require a new parser for each format, you can use eRDF with totally generic tools. Dan Webb’s Sumo is neat, but you have to configure each new format (whereas eRDF only has one).
    And once you’ve got your data into triples, you can do what you want with it: if you haven’t already, you might want to check out simile.mit.edu for some very cool tools and toys that you can easily use with RDF data.

    So what more does eRDF (initially inspired by microformats) have to learn from Tantek and co. (marketing perhaps?)

    Keith A - 18th March 2007

Comments are now closed.