Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!samsung!munnari.oz.au!sirius.ucs.adelaide.edu.au!augean!sibyl!ian From: ian@sibyl.eleceng.ua.OZ (Ian Dall) Newsgroups: comp.arch Subject: Re: Alignment on RS/6000 Message-ID: <899@sibyl.eleceng.ua.OZ> Date: 29 Nov 90 03:28:45 GMT References: <893@sibyl.eleceng.ua.OZ> <46760@apple.Apple.COM> <897@sibyl.eleceng.ua.OZ> Reply-To: ian@sibyl.OZ (Ian Dall) Organization: Engineering, Uni of Adelaide, Australia Lines: 34 In article aglew@crhc.uiuc.edu (Andy Glew) writes: >I wrote: >> Programs should core dump instead of emulating IEEE NaN and Inf handling. > >Hmmm.... what should a spreadsheet do when the user has said to divide by 0? I conceed that there are times when the IEEE defined behaviour is desirable. Basically, any interactive program which does trivial amounts of floating point are better off generating NaNs than core dumping. One such case which I ran into was in Gdb. I have a machine which *does* core dump rather than generate NaNs and Infs. It is very annoying to have gdb core dump because you asked it to print a union which doesn't happen to currently have a floating point quantity in a floating element. Don't misinterprete my position though. I never claimed that one shouldn't be able to alter the behaviour. Certainly the programmer ought to be able to specify alternate trap handlers. In the case of your spreadsheet, you probably don't want all the dependent cells filled with NaN's and Inf's any more than you want a core dump. What you really want is to enter some more "user friendly" handler of errors. As some people have said, it is possible in Suns unbundled compilers to specify either behaviour as a compiler option, which is, I think, a step in the right direction. Especially since the best option when developing a program may not be the best for the final "production" version. -- Ian Dall life (n). A sexually transmitted disease which afflicts some people more severely than others. ACSnet: ian@sibyl.eleceng.ua.oz internet: ian@sibyl.eleceng.ua.oz.au