Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!mips!apple!agate!usenet.ins.cwru.edu!abvax!iccgcc!herrickd
From: herrickd@iccgcc.decnet.ab.com
Newsgroups: comp.arch
Subject: Re: Compilers and efficiency
Message-ID: <4474.28219270@iccgcc.decnet.ab.com>
Date: 3 May 91 21:40:15 GMT
References: <9782@mentor.cc.purdue.edu> <4082@batman.moravian.EDU> <11411@mentor.cc.purdue.edu> <1991Apr28.151831.2768@lth.se>
Lines: 42

In article <1991Apr28.151831.2768@lth.se>, bengtl@maths.lth.se (Bengt Larsson) writes:
> In article <11411@mentor.cc.purdue.edu> hrubin@pop.stat.purdue.edu 
> (Herman Rubin) writes:
>>Have any language designers or hardware designers ever asked for input on the
>>variety of operations which are readily understood by people like Silverman,
>>Lehmer, etc.?  
> 
> Why do you assume that people should come and ask you (a matematician)
> about computer language development?  What makes you an expert in the
> field? 
> 
He is a customer.  Unfortunately, he does not represent a large enough
subset of their customers.

Unfortunately because he is better informed than many other customers
and all the customer base might be better served if Henry's ideas
could penetrate the language design community.

Not talking to the mathematicians let the FORTRAN design team at IBM
put the label REAL on a finite subset of the rationals.

We'll never displace floating point numbers now, but what might we
have if the game theorist had spent more time talking with number
theorists back around 1940-1945?

How many engineering juniors believe that

      DO I = 1, 50000
        SUM = SUM + DATA (I)
      END DO

computes a good approximation of the sum of 50000 numbers they measured
somehow?

Could we have realized in hardware a number system that does give us
a commutative addition?

I think there is good reason for the hardware and software architects 
to listen to the number theorists.

dan herrick
herrickd@iccgcc.decnet.ab.com