Twitter and blogging won't help those who already use it, and those who don't use it yet won't want to spend their time reading regular updates. That said, it might be helpful for those who consider using it but haven't found the time or resolve to actually do it; but these will be more interested in what newbie Cayenne users have to say than in what the developers think is the newest and greatest.
What's important is to lower the entry barrier. E.g. make Modeler intuitive to use and cover all aspects that could be reasonably modelled. (My experience, as just one data point: I toyed with it for half an afternoon and found it a bit hard to get a handle on it and on what features it actually supports. Another detail might be that the tool should announce itself with a phrase that allows people to decide what they can and can't expect it to do; for example, that it is not supposed to model everything that their database can, but everything that... well, no idea what exactly its area of expertise should be.) The documentation is actually great as an overview. It touches everything one would ask when trying to determine what Cayenne can and cannot do. It is frugal with details though. My advice would be to get Cayenne ahead. That's going to gain more followers than trying to do anything marketing-wise - the marketing that led to my current interest in Cayenne wasn't twitter feeds or blog posts, it was favorable mention in online articles. What's important is what Cayenne can and what it cannot (or will not) do. Example projects would be nice; have a web service and a J2SE application (one of each kind). Have the example projects touch every complication once: long-running transactions, distributed commits, proxy objects, optimistic update conflicts. In the famous words of Linus Torvalds: "Words are cheap. Show me the code." (I have been bitten too many times by believing some project's overhyped self description. I bet a lot of developers out there share the experience, particularly those who are in a position to advocate an architectural switch. Nothing that the developers could write will help overcome that scepticism; only working code will, and it won't convince, at best it will lower the barrier. I, for an example, still haven't committed to Cayenne; the kinds of problems that show up in the mailing list are currently making me a bit more sceptical. I'm simply not prepared to spend several person-months on an experiment that may fail, my time budget does not allow this (unfortunately, I'd love to try Cayenne out).) Regards, Jo
