Hi, I've just been assigned to a ORM mapping project (data-centric app with C# code and SQL Server database) for which the requirement is explicitly that POCO will be used, and that the ORM will autogenerate the database from the C# objects - so the C# object model is in effect king of the data.
I'd not heard of POCO before and I'm instinctively quite concerned about the idea of the database tables being autogenerated, since (a) all my experience leads me to believe that in a data-centric application, the data model is the thing that you must get right first, and (b) I'm sceptical that autogenerated data schema would be remotely optimal in terms of indexes, performance, etc.
Can anyone direct me to any links (or directly explain) the advantages and disadvantages of this approach? I'm trying to understand it better but having trouble finding much conceptual information about it.
Thanks