INFO 202 Fall 08 Blog

I’m home sick, and too out of it to do my other work, so of course I’m thinking about the tagging discussions in 202 . As Bob mentioned in one of the last few lectures, his least favorite tags in delicious are things like “web” and “toread,” which he feels are prime examples of Doctorow’s point that “people are stupid” – specifically, too stupid en masse to create useful metadata (Bob, please correct me if I’m misconstruing your viewpoint here).

I’d like to advance a contrarian approach here, which I’ve thought a lot about since this came up last year in 202. I think that the tags “web” and “toread” are the two most interesting tags on delicious, for different reasons.

The tag “web” is interesting precisely because of its evident inanity. The only thing you can tag in delicious is a website URL, which is by definition on the web, so why do people use the tag so often? Here’s my quick take on some reasons one can guess from the things people have tagged “web” in delicious:

• “web” indicates that the page has the Internet or web development as its subject. There are a lot of these pages on the Internet, and they’re likely to be popular among delicious users.
• “web” is used in conjunction with other tags to distinguish an online thing from an offline thing – so tagging something “web” and “tv” or “web” and “sketchpad” distinguishes streaming video or a sketchpad flash application from the real-world versions. Again, lots of this online.
• “web” is the top level of many people’s tagging hierarchies, so it gets used a lot.

That last one needs a little explanation, because tags aren’t hierarchical, right? Well, I’ll put out the theory that they are – but not in any strict sense; rather, in the Roschian sense of linguistic category hierarchies. If I’m tagging a page about a cat (as I so often do), I might tag it first with a “basic-level” category tag (”cat”), a specific word like “kitten” (I think Rosch calls this the “subordinate” level), and an abstract, superordinate category like “animals” (plus, of course, other modifiers like “lol”, “justhanginginthere”, etc). I’m not really thinking about it hierarchically, but I end up putting in hierarchical tags because that’s often how language works. Now if I’m tagging a lot of things, many more things will have the top-level abstraction. If this was actually a rigid hierarchical tree structure, the distribution graph would look like a power law graph – and what do you know, it often does. So the theory here is that “web” is one of the superordinate tags that gets used regularly, and because it applies to so many things it’s disproportionately popular. If I was only picking one tag, the distribution would probably favor the basic-level tag (and, probably since many people only use one tag, arguably basic-level tags like “javascript” are, in fact, as popular as “web”). But the tagging system specifically promotes using as many tags as possible, so I’m likely to pick superordinate and subordinate words as well. The tag “web” by itself may be useless – but as the top level of an implicit hierarchy of Internet-related pages, it makes perfect sense.

I find “toread” interesting for a totally different reason. Unlike most other tags on delicious, which could be intended either for the individual or for the community, “toread” is wholly personal – it says, this is something I’m interested in, but not that interested in – maybe something I aspire to read, feel faintly guilty about not reading right now, but definitely will read later. If I get to it. I use the delicious plugin for Firefox, which suggests popular tags by other users for the page I’m tagging, and every time it suggests “toread” I laugh, because it means that so many other nameless people had this complex emotional reaction, usually the same reaction I’m having at that very moment, of aspiring to knowledge they can’t quite find the time for. “Communist Manifesto” from the Gutenburg project? “toread”. Page and Brin’s seminal paper on the PageRank algorithm? “toread”. More, I think, than any other tag, “toread” makes a delicious user feel like they’re part of a community, sharing this funny personal moment with other users.

Sorry, this ended up more like an essay than I intended – I’d be interested in any thoughts.