Yeah, I never understood why most ORMs do things that way. Why the extra config file to define everything? You can infer everything you need to build an object model directly from the database schema. It's just as easy, if not easier, to read programmatically than XML. And it's every bit as descriptive.
If you're using a config file to define things, it's generally the case that it's defining the schema as well and the database is built programatically on top of it. So it's really just a matter of taste: some frameworks have you manipulate the DB schema directly and then build the ORM representation off of the DB, whereas others have you directly define the ORM representation and then build the database off of it. Either way there's no real duplication, and there's still just one place to modify if you need to change the schema/ORM mapping.
In our case we have a config file because A) it makes it easier to abstract the schema across different databases and B) we have extra metadata in there above and beyond just the schema information.
I guess that's the luxury of being a shop that builds software vs. a shop that builds open source ORMs. We can target a single DB and technology, and not need to cater to thousands of angry devs with different naming conventions and religious views as to the virtues of Postgre vs MySql.
Wrapping a relational database with an object layer is really a simple, straightforward problem. It surprising that so many smart people feel compelled to spend so much time and effort engineering overly complex solutions to it.
Well, I work at a shop that builds software, but it's installed enterprise software, which means we don't fully control the deployment platform and have to support multiple databases. Control of the deployment model is definitely a huge, huge plus of a SaaS model, but the economics don't really work out for that in our industry. We also happen to write our own (currently closed-source) ORM framework to build on top of. One of the nice things about building our own framework is that we only have to deal with the complexities that we care about (i.e. our sort of performance requirements, our sort of object graphs, the databases our customers need).
The problem with pretty much any kind of framework, ORM or not, is that eventually it becomes bloated by being flexible enough to meet the requirements of thousands of different developers, meaning that any given one of them only ever uses a small percentage of the overall functional footprint. Keeping things simple yet powerful is definitely the hardest part about it all, especially given that different users have vastly different opinions about what's simple or what power they need.
Indeed. Building frameworks for other people just doesn't sound like any fun at all. We also rolled our own thing in house, and every once in a while we start thinking thoughts about polishing it up and releasing it to the world.
But the amount of polish that you need to turn an in-house thing into an open source project is just immense. And worse, by the time you have it flexible enough to handle every possible use case, you've bloated it out to the point where it's no fun to use anymore. Pick pretty much any off the shelf ORM to see the end result of that path.
