Melissa You might find chapter 23 "The pragmatics of graph reduction" in my 1987 book worth a look. It gives other examples where CSE can be harmful. Simon | -----Original Message----- | From: haskell-cafe-bounces@haskell.org [mailto:haskell-cafe- | bounces@haskell.org] On Behalf Of Melissa O'Neill | Sent: 24 July 2007 23:36 | To: Haskell Cafe | Subject: [Haskell-cafe] Space usage and CSE in Haskell | | When advocating functional languages like Haskell, one of the claims | I've tended to make is that referential transparency allows the | language to be much more aggressive about things like common | subexpression elimination (CSE) than traditional imperative languages | (which need to worry about preserving proper side-effect sequencing). | | But a recent example has left me thinking that maybe I've gone too | far in my claims.