>On Mon, 29 May 1995, Richard Tilmann wrote:
>> Perhaps I have missed a part of this thread, but I have a concern.
>>
>> I understand the concern for seeking efficiencies in use of bandwidth,
>> but .... the use of 'common objects' seems to move away from the ability
>> to design and create unique and individualized VR spaces. I've seen an
>> increasing number of convolutions and complexities added to the original
>> idea in order to address issues of orientation, scale, color, surface
>> maping, etc. that we are almost right back to the place we started from.
>
>That's mostly a quality of implementation issue, though - if we can show
>there are cases where such caching/common objects help (I definitely think
>we can) and that it can't really be solved by other means (I think we can
>make that case too) then it should be allowed. This isn't to say
>*everyone* needs to use it all the time, it just allows it. It could
>definitely be overused, but I think the average human ego won't make
>same-ish scenes that much a danger. Besides, given the right
>tranformations you could probably look at the same teapot 100 times in
>100 different ways and not get bored. :)
Perhaps we can accomodate all sides with a HINTS phrase that is left open?
The text following the HINTS would be ignored if not understood by the
browser, but might provide valuable performance optimization if both the
space's author and the user's browser recognize the same language.
In other words, we could provide a structure for future optimization
communication between the space and the browser without specifying what
that communication may be, except, perhaps, for the basic structure.
-- Tony Godshall, Kaneohe Hawaii USA