An earlier book by one of us, [Peyton Jones 1987], covers similar material to this one, but in a less practically oriented style. Our intention is that a student should be able to follow a course on functional-language implementations using the present book alone, without reference to the other.
The scope of this book is somewhat more modest, corresponding to Parts 2 and 3 of [Peyton Jones 1987]. Part 1 of the latter, which discusses how a high-level functional language can be translated into a core language, is not covered here at all.
We focus exclusively in this book on the ‘back end’ of functional-language compilers. We make no attempt to discuss how to translate programs written in a fully fledged functional language [...] or how to type-check such programs.
I imagine this would still be useful if compilers like Haskell can produce intermediate code for what's described in the book. Not sure if they do but I have an academic paper that created a G-machine to Forth (CPU) compiler for embedded Haskell. Made me think there was a G-machine output option.
Point being, plenty one can contribute to FP in terms of safety, performance, integration, and so on just knowing backends. Esp if more people do FP CPU's.
So, the trick to answer our question is if any knowledgeable readers can tell us whether what's in this guide helps on the path to understanding that language. If so, then it's a start to working with and improving it.
http://research.microsoft.com/en-us/um/people/simonpj/papers...
I haven't read the tutorial, but the book is strongly recommended.