Trying to summarise some points that have been raised in this discussion in order to structure it for people joining in later.
- There’s interest in working on both database abstractions as well as database-development tooling
- People have feelings about ORMs
- Common database abstraction interfaces should be a basic priority
- Especially for pipelining/ batching queries
- Work on tooling and libraries bottom up: don’t start with an ORM
- There’s many efforts already in motion: no need to start from scratch
Personally I feel that higher-level abstractions (such as not having to write sql by hand, etc) are desirable. But they can be added on top of a solid base of libraries and tooling later.
I also feel that the points I just listed would make for a good guiding charter of a working group. Feel free to raise new/ challenge existing points though.