Re: Rational approximation link

From: brianjparker_at_[hidden]
Date: 2001-10-21 22:57:55

--- In boost_at_y..., Daryle Walker <darylew_at_m...> wrote:
> on 10/21/01 12:42 AM, Brian Parker at brianjparker_at_h... wrote:
> 
> [SNIP]
> > Looks good. I am undecided which is the best interface though- to
> > specify the error bound by passing in the maximum denominator as 
in
> > the above code, or passing in the relative error as in the Netlib
> > code.
> 
> Maybe we should have both.

Yes, that is my conclusion as well. It may be possible to overload 
the function based on the error term type- double or float for 
relative error and integer for max. denominator, although this may be 
too fragile and separate function names may be better.

An advantage of the relative error approach is that a default small 
value close to the epsilon of the floating point representation could 
be specified, so that the error argument could be ignored in many 
cases.

Also, as was pointed out in a previous posting, the Netlib code needs 
to have two defines updated for IEEE754 floating point- I will look 
into this in the future.

,Brian Parker.