Has a much better FFI story with JavaScript than Elm, and has a more powerful type system (higher kinder types, etc). Row polymorphism makes writing pure code a joy, the effects system is unobtrusive. I actually enjoy programming purescript more than Haskell.
However, if you want an easy migration path, there is https://github.com/pselm/core Which is a drop-in replacement of the Elm 0.18 API in purescript. You can almost directly copy your elm code to it and it will work with minor changes
Slamdata had some specific requirements with halogen that everyone may not need. And although the free monad approach is nice, it certainly is not fundamental or necessarily the best depending on the task to UI construction.
Components are these things that have a way to initialise themselves, an internal state, a way to receive messages and update their internal state, a way to compute from the internal state (ie rendering), and a way to reply to messages that are querying things (by having the message pass a callback for the reply as an argument). The only thing they have which is missing from standard oo is a way to pass up messages to their parent component.
After all that one must wonder what the point of using purescript for this is and why not just use a more object-oriented language fron the start?
I think anyone who is already comfortable with elm can look at that and be productive in purescript very quickly.
* Of course, with purescript there is going to be a bit less hand-holding. But the side-projects I see folks posting about in done with elm are done by folks that already do things on their own.
BTW my question could sound like I am hoping Phil Freeman has given up the project leadership, which isn't what I meant to say. I should have said "I was wondering whether you are mixing up ...".