Note dated May 30, 2007:

“My gosh, there is so much going on in Search this week as to not be even funny.

“BTW, note that not only does Quintura offer search for kids, but ‘Coming soon Quintura for Women…’ Can you imagine what’s going to happen when Quintura for Men hits the web?!”

However, upon revisiting the site today, I see no sign of “Quintura for Women” (although the kids’ link is still up). What a pity. It would have been so much fun to see how the tag clouds formed.

I’ve been fascinated by tag clouds, and the utility they offer by presenting multiple levels of information in an integrated and intuitively navigated way. Why are tag clouds “acceptable” and used in some contexts, and not in others? Why do I feel that a desktop-viewed Quintura-like presentation is less likely to find uptake than the tag cloud in the right column of this blog? When are tag clouds appreciated, when avoided? Which factors have the most influence: type of information? UI context? interaction means? search context?

By that last suggestion (”search context”) I mean to ask, Does it matter what kind of search I am performing? For example, when I use a tag cloud to find a post in a blog archive, there is a limited amount of data being shown to me, but that data probably includes what I need to know (what tags are available + how many of them there are, or how popular they are). The tags do not tell me anything about the chronology of the blog postings, the authors of the posts, the server location where the posts are stored, or even very much about the content of the posts. But that’s OK for the purpose at hand.

On the other hand, when I do a search on my desktop (or laptop) computer, it’s already hard enough for the search engine to figure out what I’m actually asking. (Does “rice history” request information about the history of grain, politics, or a university?) Presenting the results with a little bit of contextual content allows me (the user) to sort relevance for myself, based on the actual data. Presenting results in a tag cloud makes it harder to embed that content, making it less relevant for my use case.

On the other other hand (think Thai goddesses here), embedding the contextual data within the tag itself would partially offset the limitations of the tag cloud. Even better, embedding the path to the data and its related data would provide a means of searching within the tag cloud image itself. Enter semantic visualization.

Semantic tagging (of whichever type) describes the links and connections between different “nodes” of data. Semantic tagging has been primarily focused on providing the ability to computers to search large amounts of data intelligently — meaning, with “understanding” of the relationships within the data. However, a user interface that organizes information based on semantic links provides an opportunity for the searcher to navigate through enormous amounts of data very rapidly, according to whichever connections are meaningful to him or her at that moment. It allows for new connections, new relationships, and new meanings to be elicited through visualization of data connections. In fact, I’m not sure whether the semantic web will turn out to be more important for automated search or for human navigation of the results.

Those Quintura tag clouds are not, strictly speaking, tag clouds. They just appear to be so at first glance. They are tags leading into semantic groupings of information nodes. [I just love the way they present the visual mapping. There is something so alive about it. Which is why my husband bought me the Visual Thesaurus from Thinkmap. Not because I use a thesaurus all that much, but because I love to play around with it, and to create interesting collections of words. (Stop snickering.)]

Which brings me back around to some exploratory work I did this summer on mobile search UI. You can take a look at the early work as presented at the MobileMonday in Helsinki here.