Matthew Ingram at Gigagom blogs about an upcoming Twitter feature called Twitter Annotations. Well, it’s not actually a feature. It’s the ability to attach metadata to a tweet. This is potentially great news, since it will give us a way to add context to tweets and to enable machine-processing of tweets, not to mention that URLs could be sent as metadata rather than as subtractions from the 140-character limit. This is yet another example of information scaling to the point where we have to introduce more information to manage it. How about one of those bogus “laws” people seem to like (well, I know I do): Information sufficiently scaled creates a need for more information.

Twitter is specifying the way in which Annotations will be encoded, but not what the metadata types will be. You can declare a “type” with its own set of “attributes.” What types? Whatever you (or, more exactly, developers and hackers) find useful. Matthew cites a number of folks who are basically positive but who express a variety of worries, including Google open advocate Chris Messina who warns that there could be a mare’s nest of standards, that is, values for types and attributes. Dave Winer takes Google to task for slagging off on Twitter for this. I agree with his sentiment that Goliath Google ought to be careful about their casual criticisms. Nevertheless, I think Chris is right: Specifying the syntax but not the actual types and attributes will inevitably give rise to confusion: What one person tags as “topic,” someone else will tag as “subject,” and some people might have the nerve to actually use words for types in, say, Spanish or Arabic. The nerve! [THE NEXT DAY: Here’s Chris’ original post on the topic, which is more balanced than the bit Matthew excerpts, and which basically agrees with the next paragraph:]

But, so what? I’d put my money on Ev Williams and Biz Stone any time (important note: If I had money). You couldn’t have seriously proposed an idea as ridiculous as Twitter in the first place if you didn’t deeply understand the Web. So, yes, Chris is right that there’ll be some confusion, but he’s wrong in his fear. After the confusion there will be a natural folksonomic (and capitalist) pull toward whatever terms we need the most. Twitter can always step in and suggest particular terms, or surface the relative popularity of the various types, so that if you want to make money by selling via tweets, you’ll learn to use the type “price” instead of “cost_to_user,” or whatever. Or you’ll figure out that most of the Twitter clients are looking for a type called “rating” rather than “stars” or “popularity.” There’ll be some mess. There’ll be some angry angry hash tags. But better open confusion than expecting anyone — even the Twitter Lads — to do a better job of guessing what its users need and what clever developers will invent than those users and developers themselves.