I really like this book. I haven't finished it but I did kickstart a friend on Haskell by getting him a copy as well.
It is worth noting that the book uses HUGS examples instead of GHCI so there are some (mostly minor) differences between the example interactive code and the results in ghci. I don't remember off hand but I started the book thinking I'd just use ghci and that it would be the same as hugs, but there are a couple important differences between hugs and ghci that impacted some of the examples in the book. I stuck with ghci but to someone completely new to haskell it might be a confusing point.
I suspect this means it used top-level declarations in the REPL. Happily, with GHC 7.4[1] (released just now), GHCi accepts any sort of declaration, so it should be much less confusing.
Hallelujah! I do not understand why that feature was deprioritized for a decade, making ghci useless for quickly trying out sample code or quick ideas involving types.
+1 on this. It's a great book (not finished yet). What I like is that it's like a review (even rediscovery) of mathematical basics and how powerful they are, coupled with the beauty of Haskell. I know it sounds cheesy but _that_ is like, well, a religious feeling. Similar to what I felt when I discovered metamath.org (so wonderful).
It is worth noting that the book uses HUGS examples instead of GHCI so there are some (mostly minor) differences between the example interactive code and the results in ghci. I don't remember off hand but I started the book thinking I'd just use ghci and that it would be the same as hugs, but there are a couple important differences between hugs and ghci that impacted some of the examples in the book. I stuck with ghci but to someone completely new to haskell it might be a confusing point.