Re: Is there demand for algorithm optimization?

From: CBFalconer (cbfalconer_at_yahoo.com)
Date: 03/15/04 

Date: Mon, 15 Mar 2004 14:38:28 GMT

Trevor Barton wrote:
>
> On 12 Mar 2004 17:30:22 +0100, Dr. Michael Ulm wrote:
> > --snip--
> >
> > I am sorry if my wording was misleading. This was not what I did.
> > One improvement came from an improvement in the algorithm itself,
> > which I found using a lot of mathematics. The other, surprisingly
> > efficient, improvement came from re-writing some function. Let me
> > demonstrate:
> >
> > Suppose you need the gaussian distribution, something like
> >
> > f(x) = c * exp(- x^2)
> >
> > for some constant c. You could code it like that (if your library
> > has an exp()), but this is of course sub-optimal. Especially when
> > efficiency is necessary. After some analysis of your situation we
> > observe that we only need f for x in the interval -3 < x < 3 with
> > some given precision, say 10^(-6). Now we can try to get the best
> > method for computing this function given these restrictions. This
> > depends of course on the capabilities of the hardware. Polynomial
> > approximation (not Taylor expansion btw.) may be our first choice
> > but this problem is already non-trivial...
> >
> > Sorry, this has gotten much too technical. I stop here.
> >
> > Btw, I never claimed that I could significantly improve every bit
> > of professionally written code. I often could so far. This is not
> > because I believe that mathematicians would be better programmers
> > (they aren't), but just that I, as a mathematician can bring some
> > different point of view into it that can make a difference.
> >
> > Regards,
> >
> > Michael.
> >
> > P.S. best viewed with fixed width font :-)
>
> Good grief, do you realise that every single line in that post that
> wrapped to the next line was exactly the same length? I don't know
> that I've seen one like that before that didn't format the post by
> adding in extra spaces between some words. Certainly not one that
> long! Or is that what you were referring to in the PS?

You could have continued it by replacing didn't with did not and
also
revising the next line to:

> adding some extra spaces between some words. Certainly not one that 

-- 
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
   Available for consulting/temporary embedded and systems.
   <http://cbfalconer.home.att.net>  USE worldnet address!

Relevant Pages

  • Re: There Is No Such Reality Of Theory In Music
    ... Music does make a mention of mathematics, logical and pure, when one ... and capable of being tested through thoughtful listening. ... Music theory is an active research area of the humanities. ... wording, though, is the wording of the crackpot ...
    (rec.music.theory)
  • Re: who is the first historically recognized pope?
    ... It isn't mathematics (because the arithmetic is horseshit), ... Shame on you. ... No, it's not misleading, it's what makes him utterly wrong. ...
    (misc.writing)
  • Re: The Difference between a Set and an Element
    ... Frederick Williams wrote: ... This allegation that sets can contain concrete ... objects is misleading. ... Mathematics is abstract. ...
    (sci.logic)
  • Re: Jacksons derivation of the Lorentz Matrix
    ... branch of mathematics in itself and not just simply a once-off device ... How standard would it have to be, Pete? ... is misleading. ...
    (sci.physics.relativity)
  • Re: Any comments? (Answers to Pete)
    ... Richard wrote: ... the particular set of mathematics. ... Perhaps you should have put the equation in brackets ... That way it would not be misleading. ...
    (comp.lang.cobol)