It seems like *psci* needs to be stopped and restarted to detect changes to modules, which is very unlike Haskell and seems like a total failure in developer experience if true.

----
The whole phone book chapter of *Purescript by example* is a lot less sure-footed that the first.

The example isn't all that great with assumptions about the data and what you should be able to do with it.

It introduces Lists and Maybe but doesn't really explain how they work (and the module documentation is cryptic, really nothing more than the function signatures (*so Haskell)).

Instead it goes into how to use pointfree functions and coding style, which is amazing but still, you haven't actually explained how one is meant to get a value out of Maybe.

----
Interesting example of how design up front fails in Purescript by Examples chapter that uses a phone book as an example.

It assumes that everyone has a first and last name and makes them mandatory whereas this is neither true nor a legal requirement beyond certain territories.

----
I'm using [Purescript by example](https://leanpub.com/purescript) and its an amazingly clear guide so far.

I obviously have some background in the technologies but I thought the explanation of why you have to use both Bower and NPM was the clearest I have read to date.

That said how many package and module formats can one language have?

----
Compilation errors are not very clear. I think I had an issue with a function invocation in a arithmetic expression. Presumably I had not expressly limited my function to the Number type as inlining it seemed to work.

----
Eh. Don't really like having to hit CTRL-D to execute statements and it looks like Purescript has a shell like Haskell rather than a REPL.

----
Getting Purescript to run required nothing more than a complete reinstallation of Haskell. I have at least now gone through the process of how you do that with the pre-built binaries although how to do a pure source install is still unclear to me.