jbnv, are you serious about working on an open source version? Do you think it is worthwhile?
I've actually been trying to build a SQL/table based workflow engine with a similar design for years, and have long contemplated open sourcing what I have so far, frankly because I don't have time to finish it and turn it into a commercial product, and secondly because nobody I knew showed any interest or support.
Personally, I love this little framework I was hacking on - and I like yours too. It's a pretty generic pattern actually, but it can do so many things.
I got so far as to be able to render workflow graphs from the database, and handle branching, multiple statuses, etc. and it was modeled *very* roughly after Petri nets.
This same framework is totally reusable all over the place too, not just in ETL.
Thoughts? Anybody else interested?