Hey,
But I do expect that if we have a good idea what the optimal database
schema and usage of the feature is that you'd make a tiny effort to include the fixes that Wikidata doesn't explicitly need.
This is entirely reasonable to ask. However this particular change is not tiny, and it would cost us both effort to implement and make the change even bigger, while we're trying to keep it small. We actually did go for the low hanging fruit we did not need ourselves here, so implying we don't care about concerns outside of our project would be short-sighted. After all, strictly speaking we do not _need_ this rewrite. We could just continue on pouring crap onto the current pile and hope it does not collapse rather then fix all of the issues our change is tackling.
Instead of rewriting it using a non-optimal format and forcing someone
else to rewrite stuff again.
We are not touching this, so you would still need to make the change if you want to fix this issue, but you would not need to do it _again_. To be honest I don't understand why you have a problem here. We're making it easier for you to make this change. If you think it's that important, then let's get our changes through so you can start making yours without us getting in each others way.
Unless the ORM mapper makes it hard to use joins
It does basically does not affect joins - it has no facilities for it, but it does not make them harder either.
Cheers
-- Jeroen De Dauw http://www.bn2vs.com Don't panic. Don't be evil. --