Ayende has apologized for the way he resolved a NHibernate bug I reported. I’m glad he has changed his mind. I feel a lot of respect for people that can admit a mistake.
He also commented about the NHibernate limitations I pointed out. As he knows what the NHibernate team is working or is planning to work on, it is a very interesting read.
Regarding points 2 (old design) and 5 (hierarchical queries), it seems that they’re planning to make manipulations at the AST level, leaving the SQL generation as late as possible. This seems as a very powerful change so I hope to see that change flowing through the core to improve things.
It seems that point 3 (more events in IInterceptor) will be addressed by some changes that are being performed by Fabio Maulo. Cool!
I think I was wrong with point 1 (being too tied to the Java version) as Ayende explains.
Regarding point 4 (no paging support in some nontrivial scenarios), I think that paging when you retrieve an associated relationship eagerly should be supported. However, if the subselect problem is fixed (so not all rows from the association are retrieved), that certainly makes it less important. I was also referring to the limited paging support for SQL Server 2000. However, I’m lucky to not be using SQL Server 2000 anymore…
To sum up, the things the NHibernate team is planning for the future have made my NHibernate interest grow up a lot since my rant about it.
Ayende also mentions to try to fix bugs when you found one. I can’t agree more, however, in my personal case, I have to spend about 3 hours a day to make exercises for my back in order to be able to work with the computer without too much pain. That means that my free time is really limited and probably away from the computer :-(