IMO, it would be less confusing to compare the ML approach with classes and generics in F# (see for instance the INumeric interface, which I now see you wrote about in another post). Your approach described here using modules feels very unnatural to me.

Knowing VB or C# seems to be more beneficial than knowing an ML derivative for learning F#. So is F# ML for .Net or really just C# in functional clothes? Unfortunately, i believe too much of the latter and not enough of the former.

The purpose of the post was to attempt to implement ADT’s in a manner consistent with other ML derviatives. F# is an ML derivative implented on the .Net platform after all.

I would argue solving the above problem with clases and interfaces would be rather strange to an ML programmer. We could fall back on our C# and VB skills to do C# in F#, but would we find ourselves in the same position as C programmers who tried to write C in Java?

Your several blogs state clearly the issues I’ve been having.
The insights you bring to F# from the FP ML, Haskell and OCaml world are a balm.
I look forward to future editions.

Art