compilers@ima.UUCP (01/07/86)
[from wendyt at ALLEGRA/PYRAMID/ISIENG (Wendy Thrash)] I worked with Ganapathi's attribute-grammar stuff about three years ago at Zilog, before running off to seek (fame and) fortune at a startup. I believe we got a tape from U. Wisconsin, though we also hired Gana there, so maybe he brought it with him. Someone else eventually finished the code generator, and I'm told it went fairly easily, runs fast, and generates good code. I was never really trained in compiler theory, so to me attribute grammars were no more or less difficult than anything else -- nothing anyone with an IQ of 180 or so and one or two Ph.D.s couldn't fathom. :-) I'd also like to know what people in the know think about denotational semantics. Deciphering it seems to be about as simple as reading a German translation of _Finnegan's Wake_; is it worth the trouble?
compilers@ima.UUCP (01/08/86)
[from ] Organization: Univ of Utah CS Dept Denotational semantics is easy and natural for Lisp hackers, since it's all objects and functions. A corollary is that it gets hairy for Lisp when one starts doing destructive ops and nonlocal jumps, so formal definitions of Common Lisp are hard to find... Stoy's book is a good one, although some dislike it because it's not sufficiently precise... Can anyone refute my gut feeling that attribute grammars are a kludge? I've never found a really solid justification for their existence... stan shebs