>a strong advantage of the monadic over the relation approach is that you can actually deal with grouping in a *compositional* way such that you do not need to couple grouping and aggregation\
I created an account just to post this!
Could you talk a bit more about the difference (and similarities) between monadic and relational approaches? I am a big fan of relational algebra (as evangelized by C. J. Date). Monads, even though I don't really understand them, seem to cover the same ideas, plus more! By that I mean building up the semantics of a "relational database" by starting with Monads should give you all the power of relational algebra, plus the ability to have controlled effects, function abstraction, recursion, etc.
Again, I am not an expert on either, but that is the sense I get.
Will you publish any papers on Rx? I am very interested in learning about the ideas behind Rx, without reading through blogs or C# API docs (perhaps "Lambda the Ultimate Reactor or even Monads, the ultimate ...)? Extending "Functional Programming with Bananas, Lenses, Envelopes, and Barbed Wire" to include reactive semantics will be absolutely perfect for what I (and I expect many others) are interested in!
I liked the posts you used to do on LTU (and the interaction that goes with it), I hope you start doing those again.