I'm using Tinderbox to make this site. Tinderbox is an information management system, which to me means that I can keep information within it in ways that make sense to me --a mix of outlines, charts, and visual maps simultaneously-- and retrieve the information in useful ways: as this website or as a bibliography or directory of people, for example.
One of the reasons that I'm using Tinderbox for this site is because I'm doing a number of slightly unusual things that Tinderbox can nevertheless accomodate beautifully. I'm keeping a blog/journal, which many tools would help me with, but the entries of this blog are also nouns in a directory: hypertext authors, hypertexts, sites about hypertext, books, and so on. I want to simultaneously see entries as a chronological log and as a categorical directory of items and reviews. The flexibility to make this kind of double-vision of information is a winning feature of Tinderbox.
As for how this site is assembled in Tinderbox, this snapshot of the outline should give Tinderbox users a sense of it:
I've got the "usual" archive where entries are dumped, a hierarchical set of prototypes that set categories and export templates, and agents to gather some of the organization. Then, thanks to an article by William Cole, I got Categories working to run the directory. The only custom piece beyond that is the triage system I have going. "Featured Links" is assembled from spaces that are basically just a URL, which serves as a place to dump things that are suggested to me but which I haven't read thoroughly enough to know whether they're worth reviewing. They're available for a quick list of "stuff I've found", though. Items with are worth reviewing I'll make into a full-fledged, categorized item and put under "pending posts", which is not exported so that it doesn't take up distracting space on the blog. Items which are reviewed then go into the archive and are full-blown, published entries.