jstor however also uses some non-OSS software:

• filemaker pro, such that j. random staffer can make a "database" to track or store data (usually progress or contact information) about their area of work. just about everybody on staff has access to licensed fmp, and it gets a lot of usage.
• act is used heavily at one of the three offices, act is a desktop database jobbie designed expressly for contact management. the one office with the most diverse functional units uses this heavily to keep track of who we've ever talked to about what, when, and what the status of our relations with them and their organizations. for reasons that i am still plumbing the depths of, act didn't catch on anywhere else.

for these sorts of tasks, i think that these tools are fine: win32 and mac support are important, as these are the platforms of the people using these tools; the ability to use the same database and interface from any machine is important, so that people at different locations can see the same thing; the lightweight-ness of creating things with filemaker pro works out for a lot of folks.

not that i am considering switching everybody to something new, but are there OSS-ish tools that people have used for similar sets of requirements, or in similar situations? i'm sure we could write a program that would generate random strings of "php", "zope", "mysql", "perl", "-o-rama" and the like to describe one way of doing all this: web interface to a database. that however removes (whether actually or just in folks' minds) the lightweight-ness advantage: people need to know more details than they do with fmp, for instance, to put a simple collection of data together for their on-going use and the edification of others in the organization. or is it unfair to say that? are there OSS-ish options that retain the lightweight flavor?