thoughtcrumbs

Multiple collections and shared hierarchies

Card sorting is a powerful tool . . . yadda yadda yadda. Card sorting frustrates me, especially when I'm using it to figure out how my users mentally organize something as abstract and gooey as a website. Last year, I ran a series of card sorts with students, librarians, and instructors at the community college library where I work. The shuffling of index cards part was great, especially when people grabbed the magic marker and crossed out some jargony term like "proxy server configuration" and replaced it with something obvious like "read magazine articles online." But I felt like the Hierarchy Nazi by requiring them to put each card in only one pile. (After the first person, I gave in, and let them duplicate cards and put them as many places as they wanted.) And the software I use for rough cluster analysis of card sorts goes kaplooey if someone is audacious enough to want to put the "student computer lab" page under both "computers" and "services for students." Accustoming people to hyperlinking makes defining the "one true home of this webpage" ridiculous.

So I'm delighted with Haystack, even if it's intended for a slightly different purpose: allowing people to sort their personal files (email, to-dos, photos, MP3s) into multiple "collections," regardless of file type and much more flexibly than the strict folder system Windows imposes.

And then, I'd like to use the love child of Haystack and the Document Co-Organization project for my future card sorts. It allows people to construct a hierarchy, share it with a group, and then it generates a consensus hierarchy. Which is exactly what information architecture is about. Can we begin the eugenic software breeding project now?