Re: Re[2]: WWWlogical idea

Robert Glidden (softpres@crl.com)
Wed, 7 Dec 1994 13:29:16 -0800 (PST)


At the risk of getting in over my head (I am feeling pretty un-gurulike
about a lot of these issues):

It seems to me that VRML needs a mechanism for local objects. I agree
that the bandwidth problem is simply not going to go away any time
soon--most people are going to be using regular old phone lines for a
long time to come, plans for ATM and settop boxes not withstanding (my
gosh voice mail systems still talk about rotary dial! (If you do not have
a touch tone...))

Logging on to a VRML site (especially the same one over and over again)
might end up being an infuriatingly slow process to get access to
anything like a cool scene.

How about the SoFile node, which allows an Open Inventor file to load
another file

File {
name "myfile.iv"
}

A higher order mechanism (at the browser level?) could check for this
file locally, if it doesn't exist, get it from the remote site (or check
for date/version of the SoFile included file?). Local directory
specifications could also be at the browser level, perhaps.

These objects would not need to be standardized, but as a practical
matter they would tend to get reused (but who will standardize the _size_
of the door?).

Am I way off base here?

On Wed, 7 Dec 1994, Marc H. wrote:

>
>
> >My question: Would it be appropriate to specify a suite of "standard local
> >objects" within the specification? Then the browser would be responsible for
> >providing anything that appears on the list of standard local objects -- SLOs
> >:)
>
>
> > The HTML/HTTP folks have been looking at something similar to
> > deal with the plathora of little decorative widgets that
> > are being moved around. Presumably, in the future they
> > will just start distributing a standard library of:
> > colored lines
> > colored balls
> > graphical arrows in little boxes
> > etc.
>
> ugh. Seems like there would be no way to keep up with the jonses if
> we're talking about largely decorative objects (such as the "little red
> ball" image virus Scott's referring to in HTML). Everyone would stop
> using the standard dingbats for the same reason everyone came up with
> the little red balls in the first place -- the "standard" list bullets
> were boring.
>
> Reading the first message, I was thinking more along the lines of common
> objects (doors? books?) in a vrml space -- more analogous to HTML's <hr>
> than decorative objects. Like <hr> and list bullets, these could be
> supplanted by more decorative local versions, or left as defaults.
>
> Apologies if this is way off-base; newish to the list.
>
> /marc
>