Lists Home |
Date Index |
> 2D and 3D are essential. X3D has a 2D layer for that. The question I
> have is do you want the semantics to be in a separate class that talks
> to the 3D/2D nodes, or built into them? In X3D, you can add classes
> for that, then use the rest of the internal message machinery.
> Communication with the X3D browsers is easier via the SAI API that
> replaced the EAI.
I think that semantics on rendering engines would be optional or maybe
installed as separated layer. Take the case of 2D, when prsenting
information online one needs the additional semantics for search engines
for instance, however when preparing a XSL-FO for printing any distinction
between a <h3> and a <fo:block ...> vanishes.
I think that some people think that an underlying 2D-3D engine or format
may be fundamental. If i understand you, you are popularizing a basic 3D
format (based in X3D) with posibilities a 2D layer (of course!) where
semantic information would be stored.
I know nothing of X3D therefore let me use generic thought (of course
correct me when needed!)
Your idea may be
or something similar. I do not think that a rendering engine may be the
"host" layer. I think that host may be a "virtual machine" (VM) able to
process/transform content-oriented datument generating output.
The VM may be in a original "vacuum state" would understand the semanitcs
on the fly thanks to stylesheets linked to the original datument. This is
not like the Java-Plugin approach; it is more like XSLT or MacroTeX
approach. In a chemical notation ;-)
Datument + STS_1 + VM ==> 2D visual rendering
Datument + STS_2 + VM ==> (2+1)D visual rendering
Datument + STS_3 + VM ==> 3D visual rendering (hologram) or sculpting
(this is useful in enginnering, art)
Datument + STS_4 + VM ==> Aural rendering
Datument + STS_5 + VM ==> Search engine query
Datument + STS_6 + VM ==> symbolic computation
Datument + (STS_2 + STS_9) + VM ==> (2+1)D visual rendering + metadata
(e.g. useful when visualizing _and_ understanding molecules: e.g. charge
distributions, bond indices).
I am following this basic design in my non-XML approach, but since it is
in so early state of development i cannot offer serious thouths not
samples or code still to this list.
Center for CANONICAL |SCIENCE)