home *** CD-ROM | disk | FTP | other *** search
- Path: amd.com!usenet
- From: "Paul G. Schaaf, Jr." <paul.schaaf@amd.com>
- Newsgroups: comp.lang.java,comp.lang.c++,comp.lang.smalltalk
- Subject: Re: Will Java kill C++?
- Date: Mon, 08 Apr 1996 13:06:08 -0700
- Organization: Advanced Micro Devices
- Message-ID: <316971B0.FEF@amd.com>
- References: <3134D499.653E@ix.netcom.com> <313613B2.136E@ksopk.sprint.com> <4i7qhl$ik6@cronkite.seas.gwu.edu> <4iuhi7$fmf@sundog.tiac.net> <4iumap$mn5@hustle.rahul.net> <31582A45.3742@vmark.com> <3163C031.4FB1@esec.ch> <3164888D.2B01@concentric.net> <4kbfn8$1bu@news1.is.net>
- NNTP-Posting-Host: hqsp86.amd.com
- Mime-Version: 1.0
- Content-Type: text/plain; charset=us-ascii
- Content-Transfer-Encoding: 7bit
- X-Mailer: Mozilla 2.0 (WinNT; I)
-
- Mark VanTassel wrote:
- >
- > "Alan L. Lovejoy" <alovejoy@concentric.net> wrote:
- >
- > >Bzzzt! Not according to the benchmarks I've done. Go benchmark the
- > >factorial or fibonacci functions (implemented recursively) in both C and a
- > >good Smalltalk. You are in for a big surprise.
- >
- > You're not seriously suggesting that a fibonacci series be programmed
- > recursively, are you???
-
- Read again what he said. He is suggesting an experiment to test the speed
- of recursion; not declaring viable implementation for Fibonacci. But perhaps
- you are correct in attempting to make this explicit for other readers: DO
- NOT USE RECURSION TO IMPLEMENT FIBONNACCI for anything other than
- experimental purposes.
-
- > Have you looked at the number of operations involved in calculating, say, the 1000th term as vs a standard looping approach?
-
- That's exactly why he was using this as an example. It is easy to program,
- contains many iterations, and illustrates his point: recursion is very fast
- in Smalltalk.
-
- > I don't think this is a valid benchmark... (and I too fail to see how
- > Smalltalk can be faster than C++ except perhaps in bizarre special
- > cases)
- >
- > - /\/\ark \/anTassel (mvantassel@teambca.com)
-
- Try his experiment and you will see a set of problems for which the speed of
- Smalltalk compares to that of c/c++. You may dismiss this set of problems as
- "bizarre", but as the size of that set increases, so does its importance.
-
- Many (note that I am not saying all) c/c++ programmers claim that "Smalltalk
- is slow," and therefore dismiss it. This is wrong for two reasons:
- 1. If speed were the only measure of value we'd all
- be programming in Assembler (or tuning c compiler-
- generated Assembler).
- 2. Smalltalk is much faster than many people think.
-
- The fact is that Smalltalk can do a few things faster than c/c++, and some
- things as fast. I would also argue that most of the remaining things in
- Smalltalk are not "significantly slower" than c/c++. This point is debatable
- as it depends upon the requirement of your domain.
-
- It would be nice if we all had the resources to evaluate each of these
- languages fairly. Smalltalk is shortchanged if the programmer writes only
- c/c++-style code, just as c++ is shortchanged if the programmer writes only
- c-style code.
-
- ---------------------------------+---------------------------------------------
- Paul Schaaf http://www.amd.com |"I was thrown out of college for cheating on
- AMD, Sunnyvale CA | the metaphysics exam; I looked into the soul
- I don't speak for my company. | of the boy next to me." -- Woody Allen
-