Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!snorkelwacker!mit-eddie!uw-beaver!uw-june!pardo From: pardo@cs.washington.edu (David Keppel) Newsgroups: comp.lang.misc Subject: Re: How to make a language downward-extensible? Message-ID: <13274@june.cs.washington.edu> Date: 6 Oct 90 22:25:34 GMT References: <1990Sep24.160705.21113@newcastle.ac.uk> <9363:Sep2521:41:1290@kramden.acf.nyu.edu> <7935@scolex.sco.COM> <29047:Sep2816:51:129 <2873@igloo.scum.com> Reply-To: pardo@june.cs.washington.edu (David Keppel) Organization: University of Washington, Computer Science, Seattle Lines: 30 >stephen@estragon.uchicago.edu (Stephen P Spackman) writes: >>We already trust the compiler to do much of our optimisation. Making >>that compiler smarter, possibly interactively, is what we want to do >>next. In article <2873@igloo.scum.com> nevin@igloo.UUCP (Nevin Liber) writes: >[But does the person really know the ideal mapping to each machine?] I think the goal is that the compiler discovers ``Boy, I could move this common subexpression out of the loop if I only knew that it wasn't an alias for this thing over here'' and then queries the user to see if that is indeed the case (presumably the user says ``yes'' or ``no'' in a way that the compiler won't ask again next time). >[But won't that slow compilation?] Only the phase with optimization turned on, and only when you have it in ``query'' mode. >[Interactive might be 5-10 times slower -- any real results?] How fast can you type a response ?-) ;-D on ( Amortize compiler costs? Nah, depreciate! ) Pardo -- pardo@cs.washington.edu {rutgers,cornell,ucsd,ubc-cs,tektronix}!uw-beaver!june!pardo