For the past week or so, I've been looking into the concept of "folksonomies". These are, essentially, grass-roots classification systems. The taxonomy is flat, undefined and ad-hoc; the emphasis is on increasing numbers of "tagged" sources which will in turn provide authority despite inconsistency in tag usage. The emphasis is most definitely not on consistent or structured "tag" usage.

I've been thinking specifically about how this sort of thing could be utilized in realm of exploiting biblioblog content.

One very interesting service is, of course, Technorati. Technorati have followed the lead of del.icio.us and Flickr, incorporating this concept into their site that monitors blog post inter-linking via RSS analysis. For example, check out the Technorati page on Books. Technorati are actually pulling data via tags from Flickr, del.icio.us and Furl. They've added the ability for bloggers to add tags to posts; these are then aggregated. Technorati also (based on a brief read of their documentation on their tag usage) assume that blog categories are the equivalent of tags. So if your blogging software automatically submits your RSS feed to Technorati upon an update (most do, including dasBlog! Technorati just takes a ping and then grabs the feed), you're effectively using this service of Technorati. You can, of course, add tags (Wayne Lehman of Better Bibles Blog has begun this in the past week or so). As folks who tag their posts see which categories are most popular, they can tailor their tag usage to mirror the accepted form of the topic being discussed. For example, if there are tags like "Greek", "Greek Grammar", "Grammar" and such; perhaps one tag will win out in a "majority rules" sense. Those more popular tags are then used to describe content (hopefully accurately). Sort of an Adam-Smith-like "invisible hand" effect.

There are a few services that attempt to aggregate folksonomy categories from a more academic perspective. One service is CiteULike. Check out the CiteULike category for semantics. Even though these are somewhat narrowed down from the masses, you'll note that "semantic" is more from the perspective of folks thinking about the "semantic web" -- and not primarily works on semantics from a linguistic perspective. Also, most of the articles linked on CiteULike (at least that I've actually followed up on) link into online services that assume a connection from within an institution. Most of the links I've followed end up at something called ingentaconnect (here's an example from the semantics tag, Multigrade Predicates) which is unavailable outside of subscribing institutions.

Another service similar in nature is Connotea. My initial impression is that CiteULike has a broader base but that could be due to simply doing a mass include of bibliographic information from existing services (such as ingentaconnect). I was unable (with a few clicks, I haven't really banged on it) to find an article I'd actually want to read from Connotea. Both of these services are more like del.icio.us/Furl for scientists or academic users.

I'm wondering two things. First, is there something similar to either Connotea or CiteULike designed for biblioblogdom? I don't think so. Heck, I don't even know if that would be a good idea. But I thought I should ask the question in case I'm missing it.

Second, are there any views of tags/categories from biblioblog posts? Seems like a natural thing for Zeth (is that the person's name?) over at biblioblog.net (which seems to be dead as I write this post) to play around with since that site aggregates biblioblog content already. He could grab the categories and use those as a basis to keep track of topic counts/etc. Though that does imply growing indexes over time and supporting such things, which he may not want to do.

Anyway, just a few thoughts as I think through these sorts of issues. If you actually use sites like CiteULike or Connotea, drop me an email or post a comment about your experiences with them.

Update (2005-06-24): Joe Weaks from the Macintosh Biblioblog posts a comment reminding me that he discussed CiteULike a few months back. How quickly I forget these things.

Update II (2005-06-24): I really am late to the game here. Bob Pritchett blogged folksonomies just over a month ago. Bob points to an article by Clay Shirky that is worth reading if the topic interests you. The comments thread on Bob's post point to an article by John C. Dvorak. If Dvorak writes on it, it's usually worth at least skimming.

Update III (2005-06-24): And, of course, while we're on the topic — let's not forget ISO's Topic Maps. This is defined in ISO13250, released in 2000. Topic Maps: Information Technology -- Document Description and Markup Languages. (link is to a PDF doc). Here's the Reference Model. On ease-of-use (meaning users/authors marking up things with tags) informal folksonomies beat topic maps just about every time. But that doesn't mean topic maps aren't useful. If you google 'Topic Maps' you'll find some cool stuff.