0DBpedia's database of Wikipedia got me thinking a lot about how I go about finding information (I think all the time I've been spending with librarians lately is also influencing my brain). There are three ways. The first two are pretty common.

Searching

0Searching is when you pretty well know what you are looking for, and have some good key to get at it. For example, when I know I need to find a particular book, or a number of books by a particular author, I've got a well-defined key to get me on my way: the title or the author's name.

0The nice fields given in the library's MARC records or the particular fields for most journal searches are designed for this. RDF predicates in the Semantic Web and SPARQL queries are also good on searching, but more on that later. Searching is also sometimes a hard thing for first-year students to wrap their heads around because they've grown up in the Google world of...

Browsing

0Browsing is when you have a rough idea of the kind of thing you are looking for, so you get yourself to a likely general area (either physical or virtual) and start poking around. This, I think, is primarily what Google offers, even when you really want to do a search. That's what makes the difficulty for first-year students--searching and browsing are the same thing because research all-too-often equals "Google." That particular struggle continues.

Serendipitizing

0Here's the odder one--letting (and helping) serendipity to happen. This is when you have something stuck in your brain and it won't go away--you've been searching and browsing about it for days/weeks/months/years (ah, the good ol' days of dissertation writing!)--and then something kinda random happens and you discover a fantastic source out of the blue. This is what conferences are good for. Not necessarily the actual sessions and conferences, but the chats in hallways, elevators, restaurants, etc. that are part of conferences. That's where good serendipity often happens. And, I believe, that kind of serendipity can be fostered and encouraged. Becoming a regular at a good coffee shop, pub, or other hangout does wonders for bringing serendipitous discoveries. I think that, in many ways, universities would do well to develop serendipitizing as part of the life of the mind.

0A key element of serendipitizing is that another person has a key piece of information that you would have searched for, if only you knew to search for it. Del.icio.us goes a long way toward this by letting you see how others have tagged a resource. Others' tags could become or lead to a good serendipitous discovery.

0So if serendipitizing involves a key, like searching, and something random, or at least informal, I wondered whether the Semantic Web could become a good tool for serendipitizing. Talkdigger is already an outstanding example of this.

0To take it in a different direction, I looked at tags. Since tags give a guide to peoples' interests, I asked whether I could use the tags (categories, comments, subjects) in RSS (2.0) feeds to generate some unanticipated connections using the Wikipedia info from DBpedia.

0Success was mixed, at best. As with all good experiments, though, much was learned through the limited success.

0Here's a screenshot of what it spat out regarding an old post about the 325 Factlog, a project to document tons of info about the ancient world. I'd tagged it with "Rome". So, the query on DBpedia brought be back some images (foaf:depictions), some articles, and Wikipedia categories. Fairly tangential connections to the actual post, but perhaps that's a part of serendipity.

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0 

0

0Here's some more of it. This is picking up from the tag "Drupal." To someone who knows nothing about Drupal, this could give some more useful background.

0 

0 

0 

0 

0 

0 

0 I'm kind of intrigued by this one from Jim Groom's blog--I would never have guessed that there was a Wikipedia category, "Articles with weasel words". There might be an insight there about "Web 2.0" that that category could help hunt down.

0

0In lucky circumstances, comparing tags to Wikipedia's entries and categories made something mildly interesting pop loose. Overall, though, the key discovery is probably something I should have known all along: tagging and categories in blogs are idiosyncratic to the blog. That is, what's useful as a tag/category within a blog doesn't get far outside of that blog.

0For example, trying a number of other feeds I got a lot about the military rank of "General"--all those blog posts categorized as "General". I also noticed that in many cases tags or categories missed the key words from the post titles. Other categories proved broad enough that the hits from DBpedia got too far away from the posts topic to seem useful.

0So sometimes categories are only good for an idiosyncratic site organization, sometimes they provide useful tags that could be used elsewhere. Often, both are combined on one blog, or on one post.

0Looks like tags or categories won't get too far (at least in the fairly anemic application I built) for serendipitizing with the semantic web. But maybe there's still room to explore. Flickr's machine tags might prove more fruitful for this kind of approach.

0Something that could be very useful, though, would be a mashup between the categories on a blog and Wikipedia categories. That would make a fast tie between a blog and lots of RDF data, and make blogging categories correspond with a broader set of information.