home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.arch:10647 comp.lang.misc:3607
- Newsgroups: comp.arch,comp.lang.misc
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!wupost!spool.mu.edu!news.nd.edu!mentor.cc.purdue.edu!pop.stat.purdue.edu!hrubin
- From: hrubin@pop.stat.purdue.edu (Herman Rubin)
- Subject: Re: Hardware Support for Numeric Algorithms
- Message-ID: <BxLupy.M9K@mentor.cc.purdue.edu>
- Sender: news@mentor.cc.purdue.edu (USENET News)
- Organization: Purdue University Statistics Department
- References: <1992Nov11.100555.4706@smds.com> <721539025@sheol.UUCP>
- Date: Thu, 12 Nov 1992 13:27:33 GMT
- Lines: 37
-
- In article <721539025@sheol.UUCP> throopw@sheol.UUCP (Wayne Throop) writes:
- >: From: rh@smds.com (Richard Harter)
- >: Message-ID: <1992Nov11.100555.4706@smds.com>
- >: What Bob is telling you is that there are problems
- >: which are so computationally demanding that computational efficiency is
- >: far more important than standard programming style guides.
-
- ...................
-
- >: There is a price for programming for clean code;
- >: in most cases the benefits far outweigh the price; sometimes they don't.
-
- There is a price for programming in what the languagists think is clean code.
- Nobody except a computer programmer can read it at all! Now I received some
- email from someone about the algorithm I posted in which he asked for another
- copy of the code where the control-path was "clear" and not defined by breaks
- or gotos. Now I do not know how to do that, but analyzing breaks is not that
- simple for a human, either.
-
- >True, but I'm more sympathetic with the Jon Bentley school of
- >optimization, where you transform clean, correct code into arcane,
- >incomprehensible code by known-safe transforms, and RECORD THE
- >TRANSFORMS. For example, the chapter on code tuning in Programming
- >Perls, especially the case study of the binary search. Or, if I'm
- >remembering rightly, Bentley's book Writing Efficient Programs. Of
- >course, the shortcoming in this context is that the central focus isn't
- >really numerical problems, but I regard that as relatively minor.
-
- We still have the problem of what is comprehensible to whom. And what
- is one supposed to do if the hardware operations wanted are not even
- expressible in the language? Sure, one can simulate them, but nobody
- will be able to read THAT code.
- --
- Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
- Phone: (317)494-6054
- hrubin@snap.stat.purdue.edu (Internet, bitnet)
- {purdue,pur-ee}!snap.stat!hrubin(UUCP)
-