[MUD-Dev] Breaking down the walls

Nathan F. Yospe yospe at kanga.nu
Mon Jun 24 02:59:46 New Zealand Standard Time 2002

Hans-Henrik Staerfeldt <hhs at cbs.dtu.dk> said:

> I implemented circles-only. The question is wether the _exact_
> shape makes much difference in a textually rendered
> environment. If i wanted to make computational shortcuts and
> implement general polygons i would circumscribe them with a circle
> and then only test the polygon if the circle contained the point.

I do similar things with spheres for my server core and graphical
client.  The real key is the intersection between two spheres and
a set of vectors... the Center of Mass and Center of Space points
are obvious enough, and the Extension of Space defines one of the
spheres, but there is also an effective space sphere used to find
the extension possible for a non force-affected object, given the
movement of the CoS around the CoM.  Add to this Acceleration and
momentum, angular and linear, and Jerk, Jolt, and precession, and
it could get very complicated without these factors.  (Also, when
a computation is performed, it is performed on a collection, with
a potential vector set for delta/delta dot/etc on the EoS.)  What
you gain is the ability to do most computations on points, and to
only perform further computations (and Coordinate Transforms, the
mass application of which are the mainstay of fast evaluation) on
objects that interact.

A tumbling cylindrical object still looks like a sphere to any of
the objects it interacts with, until physical intersection with a
part of that sphere occurs.  (Really, a point, until it meets the
sphere and chances on an occupied part of the sphere, or a region
between an occupied part and the center of the sphere...)  And to
the environment, it looks like a statistical amalgamation of what
passes for simplified fluid dynamics for the structure...

Of course, that's assuming an inert object.  A person who blows a
lot is (barely) going to have to count as non-inert, because some
of the mass (air) becomes fluid, and can't be reasonably used for
the collection object.

(Yes, for those familiar, I've given up on treating fluids as the
virtual collective objects I was, and instead have diffusive sums
and pooling, blending, and statistical diffusion on regions, from
the containment boundaries.  The world is containers bound by and
contained by other containers... I don't support intersecting but
non containing containers, causing pass-through to become a split
and transfer and expire operation for containers at boundaries.)

> (code available on request)

(code not available... sorry, but I have plans for it...)


Nathan F. Yospe -  Programmer, Scientist, Artist, JOAT with a SAK
yospe#kanga.nu   Home: nathanfyospe#mac.com  Work: nyospe#a2i.com

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list