GMP v MPFR

[Linas Vepstas]

On Sat, Apr 08, 2006 at 01:09:39AM +0100, Jim White wrote:
> 
> there is still one particularly
> attractive feature of MPFR - it's rich elementary
> function set.

Ah HA!  So you say that MPFR supports some elementary 
special functions ... this is not a claim that the web 
page actually makes. If this is true, then it would make
MPFR interesting (to me). But there is no hint of this
support on te MPFR website.

-----
Well, spurred by this email, I dug deeper, (a lot deeper)
and found the special function support. I'm rather annoyed
that after crawling over MPFR and other related pages for
an afternoon, and engaging in this odd exchange of emails,
that I was unable to find the special function support
until this email clued me in.

Can I advise that someone update the MPFR web page to 
actually list what it supports? i.e. to prominently
proclaim something like the following:

  * MPFR supports precise rounding for arbitrary 
    precison floating arithmetic, using a programing
    interface tht is similar to the GMP mpf_* interface.

  * MPFR provides a library of elementary special 
    functions, including the trig functions (sin, cos, 
    etc. and inverses), exp and log, gamma, erf, zeta
    all for real args (no complex arg support).

  * MPFR provides ?? (what else isn't obvious unless 
    one crawls through the docs??)

--linas