I'm still pretty new to Twitter, but it feels pretty familiar to me. Part of the reason for that is that some of my Facebook friends have been parallel posting their status on Facebook and on Twitter. This mostly annoyed me, because Twitterers update their statuses more often than facebookers, and a lot of the Twitter vernacular is totally inexplicable when viewed on Facebook. On Twitter, by contrast, I find that I'm annoyed when people that I follow, but don't really know, mix details from their personal lives into their otherwise interesting Twitter streams. For example, I follow dchud because I know that he will throw out some very interesting ideas. (He was the very first person to follow me on Twitter; I was the very first person to comment on his blog way back when). But I'm not really interested in his reports on the Washington Capitols. Somehow I find that Facebook is a much better place to get to know details like that- if you friend me there, you'll find that I'm a rabid fan of the Philadelphia Phillies, and I won't mind it if you update me on the triumphs of your Columbus Blue Jackets. We both knew they would lose eventually.
The thing that intrigues me about Twitter is that it does so little so well that it's really a lot easier to fix the problems that it has. The past two days I've been writing about the challenges of propagating vocabulary (and grammar for that matter!) for use in the semantic web. Yesterday, Google demonstrated one way of propagating vocabulary- be big and powerful and just tell the world what vocabulary to use. Ian Davis called Google's approach to implementing RDFa "a damp squib" which is what Americans would less colorfully call a "dud" or a wet firecracker. He lamented that Google had chosen to use their own limited vocabulary rather than adopt vocabulary already in use. R.V. Guha, who I mentioned in yesterday's post, commented on Davis' blog that we shouldn't judge too soon what Google is doing. A lot of us are hoping that "igniter fuse" will turn out to be an apter pyrotechnical analogy.
The other strategy for vocabulary propgation is based on community-based collaboration. In my post yesterday, I complained that it was hard to find vocabulary that I might use to attach an ISBN to a resource. In contrast, Twitter, together with the accessories that are built around it, seems to enable rapid propagation of vocabulary and grammar. So back to my complaint about how Twitter streams seem to annoyingly mix tweets of varying interest. One way that people deal with this is to use multiple accounts to organize their tweets in different genres, the same way you might want to have a business email and a personal email. Perhaps a better, more flexible way to address this would be to adopt a special hashtag to signal that a tweet is not a product of one's brilliant intellect, but rather just a status message about "my personal life" (#mpl). That way, your mom can easily filter out your irrelevant work stuff and and your boss can filter out your irrelevant personal stuff. Anyway, if you think this a good idea, see if you can help propagate the use of #mpl (let's call the idea #mplIdea). If you don't think it's a good idea, or if there's some better way to fix this twitter deficiency, leave a comment. Let's see if we can demonstrate the power of the collaborative approach to vocabulary propagation.
The adoption of RDFa by Google and their centralized approach to vocabulary may be a turning point in the first stage of the semantic web- that of using the web to aggregate data. I think this approach is not going to take us very far. We need to start building the second stage of the semantic web. We should be thinking about collaborative intelligence rather than about accumulating distributed sets of data. I don't expect that machines will be able to come up with ideas like #mplIdea, but I do think it's reasonable for machines to be able to help us judge wether ideas like #mplIdea are inspired (and should be propagated) or whether they're just stupid.