All things considered, I'm quite happy that I've not yet released Bermuda (I wish use.perl supported tags!). Currently I'm working on a conceptual problem dealing with separating the generation of data structures with the serialization of said structures. By making the separation more explicit, much of my code has simplified quite a bit and I now see an issue with some class names being confusing. Had I already released this code, I would be much more concerned with others using it when I still want to break the API. By talking about it, I've gotten a lot of excellent feedback about design issues, but even when I haven't received feedback (hint, hint!), merely talking about it has clarified issues for me.
The downside, however, is obvious: who cares? Aside from a couple of people telling me in email that they're very interested in this project and are looking forward to its release, short of releasing code, it's very difficult to tell if it scratches anyone's itch but my own. Programming in a vacuum lets me do exactly what I think I need to do (that helped when I started the TAP::Parser project), but my needs and yours might be radically different. Does Bermuda sound even remotely relevant to what you do?