Re: LANG: VRML 1.x Binary Format Proposal

D. Owen Rowley ([email protected])
Wed, 31 May 1995 08:41:08 -0700


>A few comments on Mark's comments on my comments on his proposal (say
>*that* five times fast):
>>[ I am not clear on this (unit == 1 meter) but have thus far kept quiet to
>>the whole list about this. I have - after two years of thought - determined
>>that millimeter == 1 unit resolution is almost always the best choice,
>>because most values are then integer quanities, and motion deltas (which
>>will become important in VRML 2.0) can be given as byte values, rather than
>>IEEE floats. This utilizes 4x less bandwidth for the same information.
>
>I see. Hmm; we need to resolve this (in some fashion) fairly soon,
>because it *does* become critical as soon as we start hooking pages
>up to each other -- we need to make sure we're all on the same scale.
>(Or, alternatively, that we have some way of scaling worlds.)

I think it would be nice to have that sort of chorus lined up, but in actuality
there are plenty of issues like this already.

some people like bright lights, some people like dim lights, some people
like to drive, other people want to be driven, lets face it people are
perverse.

And theres no accounting for taste
>
>I can see the 1 unit == 1 millimeter argument, and I will admit that
>I like avoiding floating-point math when possible. *However*, if that's
>the case, the 16-bit format still has the precisions a bit off. With
>8 bits for the integer part, that means that we only have a range of
>about a fourth of a meter, which is a bit too little -- we need at
>least a meter or two, I think. (Plenty of real-world objects are a
>meter in size, although not too many are more than four meters.) We
>also don't need a full 8 bits for the fractional part -- that would
>bring our resolution down to almost micron level, which is *way*
>smaller than we need.

I have thought for some time that the centimeter was a good compromise.
for one thing the millimeter is appropriate for one to one mapping of our
mean visual acuity (sp?) but we will be not be using 1:1 that often. More
likely our display windows will represent things at about ten percent actual
size.

>>For the 14.4 modem user who's downloaded their copy of
>>WebSpace/WorldView/VRML browser-of-the-month it will be the most important
>>feature of VRML.
>
>Okay, I'll take your word for it here. I still don't have a browser
>(Tony, TGS, anybody got a Windows version ready to roll?), so I haven't
>experienced the agony for myself yet. Sounds like I'm going to be
>mighty happy when my company upgrades to 256K next month, though...

the agony is downloading a few megabytes worth only to find its not
standard VRML!

LUX ./. owen
D. Owen Rowley - [email protected] - http://tcc.net.org/~owen
Organ Grinder: The Community Company LTD. http://tcc.net.org/~tcc
What this industry needs are good "people-oriented" cyberspace applications
- Because the world isn't just 3D -