Big Fractal Tangle: solitaire saves the semantic web

My wife plays Solitaire incessantly. She's got a PhD and a professor's schedule, but nearly every time I look over at her laptop, she's reflexively sorting cards into piles, playing Freecell or some variation. She tells me it relaxes her, which seems reasonable enough, even if my idea of relaxing involves a dark room and a soft pillow.

She's not alone. Solitaire is easily the most popular computer program of all time. More popular than email or web browsing, Solitaire is often the first program people use. I used it to teach my mom how to use a mouse. It's a great example of drag-and-drop. Its UI directly maps to familiar items, making it the quintessential example of transparent usability.

So why do people keep playing it? Why do so many spend so much of their down time playing this simple, repetitive, game? Well, as best as I can figure, there's a pleasing comfort in its familiar patterns, in its meditative single focus, in its quiet sense of mild accomplishment.

There's an important lesson here, one that may well help the semantic web really take off. Months ago, I wrote that the stumbling block of the semantic web was motivating regular people to annotate their data. Given that most people keep all their email in their inbox, how do we convince everyone to tag their stuff on a regular basis?

I think we've found the answer. Here at Immuexa, we've been struggling for months to give Tidepool the reflexive mindfulness of solitaire, hoping to make the activity of tagging a pleasing and meditative habit.

And if that don't work, our instant messaging mojo should do the trick. More tomorrow.

Hmm, I'm not sure about getting people to annotate their data is really /the/ stumbling block. After all, FOAF profiles aren't annotated data, they're descriptions of the world, top-level data in themselves. If you get someone to make a profile of themselves, you've got the data.

But I'm certainly looking forward to hearing what the solitaire angle is. I do find hacking at RDF/XML in a text editor often lacks reflexive mindfulness ;-)

posted by Danny at May 12, 2004 06:09 AM

In [http://www.objs.com/workshops/ws9801/papers/paper056.html], Ora Lassila says

"Metadata is "data about data" and can be thought of as (possibly formal) descriptions of some other data (which may or may not be available on-line)".

If you can accept human beings as being "data", then foaf:name is metadata! That's a philosophical stretch, though :D

Though you could always play it as statements referring to a certain URI - the URI is data, of course (which happens to correspond to a concept - you), so the foaf:name relation is metadata.

But I agree - it's not a good term! "Metadata" is also a trademark - we should really use "meta-data".

Regarding the topic at hand: the big stumbling block for the SW IMO is Carroll's production paradox. In this situation, it can be phrased as "people won't spend 5 minutes sorting their email now, even if it saves them 20 minutes finding it later". Meta-data can be enormously enriching, but it needs to be automatically generated where possible, and easily created the rest of the time.

If Timothy's found a good interface for doing it, then I look forward to seeing it!

posted by Rich at May 12, 2004 04:08 PM