A little more explanation ...
I used acs-workflow to encapsulate the processes I describe as being fairly Sloan specific, but ... the result convinced me that acs-workflow was too complex to use in everyday cases.
So Lars and I sat down and talked about workflow last June in Amsterdam. I thought it would be one of those talks where I'd politely criticize acs-workflow while taking pains not to hurt Lars's feelings, but it turned out he'd reached the same conclusion on his own and we had a lot of ideas in common as to what an easier, simpler workflow package would look like. We agreed, for instance, that Finite State Machines would suffice for the vast majority of workflows we'd use in OpenACS and would be easier to work with for the majority of programmers, most of whom don't know Petri nets. PL/[pg]SQL-only callbacks were, we agreed, a mistake. I could go on but let's just say it was a very fruitful meeting.
Anyway ... the events package is messy for that, among other reasons which I won't go into here. I'm not interested in releasing it in its current state because frankly, it just sucks too much and I don't like airing my sucky code in public, just my not-so-sucky code.
