> Thanks for your comments. I am not quite sure what to make of > them all, but perhaps a point of reference which would help us > understand: what are you comparing Cayenne to? Hibernate? > Something else? No ORM at all?
I'm being bitten by Hibernate. > I ask, because promoting Cayenne seems to fall into two categories: > 1. Cayenne is a more suitable tool for the particular task than other ORMs, > 2. You'll want to this this ORM thing instead of putting SQL into your code. > > They are quite different audiences for any messages we are trying to get out. Actually, I'm both audiences. I do have a lot of bulk SQL in my code, and I'd love to get rid of it - mostly for cross-database portability reasons, but also because dealing with the data at both the SQL level and the ORM level has turned out to still be an impedance mismatch. I doubt that and ORM is good enough at bulk SQL for productin use, though. It would require a lot of database strategy intelligence in the ORM layer; Hibernate's HQL comes close (but fails from lots of deficits in the details, so it's far less useful than it could be). The difference I meant to highlight is that favorable mention from users is going to be far more convincing than blog posts from devs. When selecting an ORM, I like to see testimonies from three kinds of people: 1) Those who are just beginning to use it. This allows me to judge the height of the entry barrier, and estimate how hard it will be to convince colleagues and superiors to give it a try. 2) Those who are in the process of learning it. What are the barriers that they hit? How much help to they get, how much help did they need? This allows me to give my superior a rough estimate on how long and painful the transition phase will be (whether transitioning from JDBC or Hibernate). 3) Those who are in the maintenance phase. Cayenne is already in regular use for them, so what's their 20/20 hindsight on it? Where are the limits, did they need to push the envelope, how much did the envelope-pushing help? This will allow me to give a rough estimate of long-term benefits. Devs can't answer any of these questions, not easily anyway. Interviews might be the way to do. Regards, Jo
