[boost] Re: Standardization of quaternion,octonion,andspecialfunction libs?

From: Hubert HOLIN (Hubert.Holin_at_[hidden])
Date: 2002-07-01 06:37:27

Somewhere in the E.U., le 01/07/2002

        Bonjour

David Abrahams wrote:
> 
> Given the significant positive momentum towards accepting uBlas, mightn't
> it make more sense to use its expression template approach than to
> integrate with PETE?
> 
> -Dave

                So far, I have not been able to find time to realy look at uBlas (and
thus can't honestly vote on it), though I eagerly hope for its inclusion
into Boost, as we truly need some form of "standard" matrix library
(here's hoping it will play nice with TNT...). I was therefore unaware
that it had some expression template mechanism.

                The work I had started with PETE is older, but since PETE seems to
have been stopped dead whereas uBlas is alive, it does indeed make sense
to use uBlas there rather than PETE (but that also means I have to start
from scratch again, so it will take longer).

                At any rate, I will need to use uBlas to properly incorporate the link
with rotations, which are currently only an example in the quaternion
library using an ad'hoc matrix type.

> From: "Fernando Cacciola" <fcacciola_at_[hidden]>
> 
> >
> > From: "Hubert HOLIN" <Hubert.Holin_at_[hidden]>
> >
> >
> > > Somewhere in the E.U., le 28/06/2002
> > > [Slowly getting back to speed]
> > >
> > > "Powell, Gary" wrote:
> > > >
> > > > Don't forget the game industry. Turns out quaternion's are used by
> > > > people who do 3D game engines. They would like/use this library.
> > > >
> > > > The only question I have is whether the current library should use
> > > > expression templates to help the compiler eliminate temporaries. I
> > > > suppose a well written specification wouldn't preclude that as the
> > > > implementation.
> > > >
> > > >   -Gary-
> > > >
> > > > PS
> > > >  Very few game programmers seem to know about boost yet, but that is
> > > > changing.
> > >
> > > [SNIP]
> > >
> > > Bonsoir
> > >
> > > Actually, I have started investigating the possibility of using my
> > > libraries in conjunction with PETE (http://www.acl.lanl.gov/pete/),
> with
> > > the aim of seeing quaternions and octonions as containers. There are a
> > > few technicalities, but overall it should work, and thus eliminate many
> > > temporaries (at perhaps the price of explitely coding the intent). I
> > > intend to add examples to that effect to these libraries (perhaps not
> in
> > > the current revision, but likely the next).
> > >
> > Last time I checked, PETE didn't compile with BCC55, so if you do that,
> > please allow the user to switch ETs on/off just like it is done on uBLAS.
> >
> > Thanks,

                The initial plan was to provide additional files, and to use PETE
(which has its own copyright issues) only if available, via additional
"#include"s. There was to be no overhead or additional compilation
problems for users of the library in its basic form.

                To use uBlas instead, well, I do not know yet.

> > Fernando Cacciola
> > Sierra s.r.l.
> > fcacciola_at_[hidden]
> > www.gosierra.com
> >
> >
> >
> > _______________________________________________
> > Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
> 
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

        Merci

                Hubert Holin
                Hubert.Holin_at_[hidden]