Steve,
In a message dated 98-05-05 10:46:28 EDT you respond:
> > ... First priority should
> > probably be a general strategy which would allow any such system to deal
> > efficiently with both shape and color ...
>
> As an extension to defining color, how about material composition &
> properties
> which may vary discretely or continuously throughout a part?
>
> Does anyone care to suggest potential applications? (Some have already
> been
> demonstrated using processes like MD*.)
Good idea. One of the obvious applications would be in medical imaging. It
would be good to be able to slice the data at any level and "print" the layer
with all the internal information available - as for MRI or CAT data.
> > The STL format was apparently designed to allow for a facet attribute
such
> as
> > you describe. This would seem the logical approach for relatively simple
> > coloration - in which different models or even different facets of models
> are
> > built in appropriate colors. It would not, however, allow for complex
> > variation within facets.
>
> Consider "Texture Mapping" as applied in the Computer Graphics
> community. If
> the only concern is surface color, it should be possible to associate a
> bit-mapped
> color defininion to each planar facet.
That was my first thought, but it seemed that since facets must subject to
manipulations, such as subdivision for better "fit," adding the color info
could be a bit much. I must defer to you and others with more software
expertise regarding whether something on the order of a 300 dpi color image
would be too cumbersome to add to facet definition and carry along through
subdivisions, etc.. The concept I floated would keep such surface info
largely independent of mesh manipulations.
> > ... Accepting the
> > STL dominance as a demonstration of its power...
>
> STL is actually quite a powerless representation. It represents parts
> using a
> collection of oriented facets sitting in space without providing any
> topological relationships between them or any way to define higher-order
> geometric primitives. If a facet is "lost" during translation, it is
> not at
> all obvious from examining the file, and the format is exceptionally
> redundant
> with vertex coordinates repeated numerous times. Its dominance is much
> more
> likely due to being first to market and the simplicity with which it can
> be
> generated.
Precisely put.
> It will serve the RP community well if an improved alternative to STL
> can be
> developed & accepted before integrating color/properties into part
> definitions.
That would be fine with me, as long as the provision for other properties is
included in the alternative's design and development. Seems that we should
try to work from the "particular" (just shape) toward the "universal" (shape
and other attributes, both internal and external). The broader the range of
applications which can be facilitated, the more all of us will benefit in the
long run. We're talking about the potential of a vastly increased market.
Any thoughts on that "3D bitmap" concept - suggested by the computer graphics
work with "voxels" done, as for medical imaging? Just might be the logical
basis for the combination of internal and external information you suggested.
I don't know how practical it might be when the "voxels" are reduced to the
size range which would be required by actual users.
Norm Kinzie
(781) 444-6910
For more information about the rp-ml, see http://ltk.hut.fi/rp-ml/
This archive was generated by hypermail 2.1.2 : Tue Jun 05 2001 - 22:45:30 EEST