[ODE] Submitting documentation

Martin C. Martin martin at metahuman.org
Fri May 30 11:41:02 2003


Russ Smith wrote:
> 
> hi martin,
> 
> more documentation is greatly appreciated!

I'll do my best.

> as for the mass functions - i've been using the convention that mass
> setting functions take the density, not the total mass

Yes, and we should definitely keep those, although I think having a second
set that takes the total mass instead of density is useful some times. 
For example, when trying to model a real object, you're more likely to
know the mass than the density, and that may be more intuitive to
manipulate.  Sure, you can calculate the density from the mass if you go
through the volume calculation, but that can be awkward and error prone. 
Also, in my application, I want simulate a rectangle with "normal" width
and length but negligible thickness.  With the "total mass" functions you
can specify a dimension of zero, as long as the other two are nonuser.

> but it you want to specify the total mass
> and not the density, then you call dMassAdjust() giving it the previously
> initialized mass object and the new total mass.

Hmm, well this may do what I need, as long as I give a small size for that
negligible dimension.  So do you think it would be wise to incorporate a
second set of functions that take total mass, or just leave people to use
dMassAdjust()?

Thanks,
Martin