Re: Interest for quaternions (and octonions)?

From: Hubert.Holin_at_[hidden]
Date: 2001-01-02 18:35:34

Paris (U.E.), le 03/01/2001

--- In boost_at_[hidden], "Dean Calver" <deano_at_r...> wrote:
> I think quaternions would be a good library to add to Boost. I (like
> everybody else it seems) have my own (well several actually).
> 
> In games we use them alot for rotations and orientation etc and would seem
> to be a natural extension to the existing number sets that the STL and boost
> already have.
> I don't personally have a use octonions (I seem to vaguely remember hearing
> the term in quantum mechinics and combinatrics) but I'd like to have them in
> Boost as you never know what you need at a later date?

        OK, I'll also upload the octonions. It just needs a bit more work 
(specialisations and coding some transcendental functions), and to 
abide by the Boost guidelines (only a bit more work).

> I also am wondering about numeric classes being added to Boost, I have
> numerous times written vector (mathematical) and matrix templates classes
> and they would seem to be obvious additions (I mean relatively simple matrix
> classes rather than blitz++ complexity) ideally we would want a generic
> tensor class. In almost all mathematically complex fields (personally I've
> needed them for graphics, physics, combinatrics, numerical analysis and even
> quantum computing) the need for the more complex (no pun intended) maths
> classes often occur.

        Not to mention mathematics :-)  ("applied" or "pure"). I believe 
the situation here is somewhat complicated by several diverging 
approaches that one would love to see working together: Blitz++ (data 
types, http://oonumerics.org/blitz/index.html), TNT (which I use at 
work, data types and algorithms, http://math.nist.gov/tnt/),...

        Actually, one thing which is missing from the quaternions I 
submitted is a "transparent" link with rotations in R^3, for lack of a 
"standard" (de facto or de jure) representation of rotations.

        I do not really see the prospect (unfortunately) of integrating 
either TNT or Blitz++ into Boost, but a Boost data type with glue to 
either (or both...) would be a wonderful thing. Any taker?

> A high performance generic numeric library would be a major addition to
> Boost. With almost all modern processors today having some form of vector
> unit (SSE, altavic, 3DNow, Emotion Engine VU0) a generic library would
> potentially allow specialized versions giving massive optimizations.

        I still gripe about lack of automatic use of vector capacity, when 
available, by valarray...

> Of course, building a library to meet everybody requirements would be a
> major undertaking. I'd be inclined by taking a straw-poll of which areas
> interest people and get some discussion going about implenting a STL
> quality/style version.
> 
> Bye,
>     Deano
> 
> Dean Calver,
> Lead Engineer (PS2) Dreamland Freedom Ridge
> Mythos Games

        One potentially huge project would be to build such a numeric 
library in synch with the release of NIST's Digital Mathematic Library, 
which is scheduled for 2001/2002 (http://math.nist.gov/DigitalMathLib/
).

                Hubert Holin
                Hubert.Holin_at_[hidden]