> I’ve never spotted an instance where something has been rewritten in a different language because c# was lacking in some way.
Yes, that is what killed F# effectively - on top of not having any killer frameworks. Though to be fair even Scala is dead. Making a language popular is tough.
F# has been an unwanted child of Microsoft, it actually adopted the language from Microsoft Research. As a language, F# is _very_ good (although not as good as the original OCaml, whose features had to be cut to be able to run on CLR). It's just not first-class citizen, so not everything may work with F#. And F#'s own development and features are blocked on CLR -- instead of F# charting its own path, each feature first has to be available for C#/CLR and only then will F# developers implement it (higher-kinded types are just one example).
But it's still much more pleasant "syntactic sugar to write the CIL bytecode" than C# can ever be.
> even Scala is dead
I beg to differ! Thankfully, Scala is very much not dead. Scala 3 is around the corner, with many more features and, even more importantly, several simplifications. And the community is also much larger compared to F# (or any other ML language, like Haskell). Innovation also takes place in the realm of libraries, like ZIO for example.
And here I am happily using it for greenfield projects, interfacing with popular ecosystems and just generally getting things done!
Right now it really feels like F# is taking off. It is bundled with .NET Core, fully open-source, has a better REPL experience with Nuget integration, Ionide and Fable just had a new releases...
What killed F# was Microsoft never putting enough internal resources on it. It's been consistently an afterthought in various releases of VS, initial .NET Core support, etc. As a C#/.NET Dev for 20 years, I've tried several a few times to use F# only to find broken tooling, lousy IDE support in VS (compared to C#, etc.) Jetbrains Rider supposedly has better F# support, but given Microsoft's lack of interest and a tiny community, I still can't be bothered to pick it up. I'm more likely to learn/use Elixir than F# at this point.
FWIW I keep hearing about F# all the time. From where I sit, there is still a lot of enthusiasm around it. It does not seem dead at all. Is there not a lot of interop between F# and C#/dotnet?
Usage of C# packages from F# is simple and without issues. The other way is a bit more clunky, but frankly, you will not want to write any C# once you are on the F# side.
I have been programming in F# full time for the past 7 years and I hope that rumours of its death are greatly exaggerated. Perhaps this comes from a point of view that if something is not growing 10x each year is dead?
It's not a JavaScript framework that has an average lifespan of a couple of years. It's a fully featured, complete, very productive language that makes programmers life easier and more enjoyable.
I wonder how much of the "death reports" are also just that F# is one of Microsoft's original successful invites for open source community contribution. It was so successful that some months the majority of Github contributions to F# come from outside of Microsoft and that can appear "dead" for "Microsoft officially supports this" rather than a sign that the project is healthy and seeing a lot of community love.
Yes, that is what killed F# effectively - on top of not having any killer frameworks. Though to be fair even Scala is dead. Making a language popular is tough.