developing new search tools

Years of pedagogical research has left us with a fairly complex understanding of the different learning styles we can plan to encounter as teachers, so that we can adapt our classroom to benefit multiple intelligences. But it strikes me that when we develop new research tools, we base our design on existing digital models without critically investigating researchers' learning styles.

Over the past few days, the 18thConnect group has been thinking about the next steps we will take after fine-tuning our OCR. Michael Simeone mentioned his desire to develop an open-source search visualization tool, so that we can graphically see what it means to have, say, 427 hits returned on a search inquiry. This got me thinking: what if, instead of just making an interesting new graphic a part of our search function, we let users choose different search "apps" or interfaces, depending on how they would like to approach their digital tools. What if, down the road, we could make our search algorithm transparent so users could tinker with it?

This stream of questions came on the heels of Tarleton Gillespie's recent visit to the UIUC campus. Over a casual STS lunch, he discussed how people-- especially humanists-- approach search functions like they are cut-and-dry, as if no other options exist. Perhaps pop-culture understandings of the word "algorithm" make search algorithms seem like an abstracted Truth, rather than one of many possible models that reflects human values in the way terms are weighted and negotiated.

I'm wondering: has anyone out there been working on search visualization tools, or on the idea of making an adaptable search interface? Any references to texts, sites, or potential collaborators would be greatly appreciated :-).