Re: Browser Observations

Check the nesting of your curly braces... :-)

You have the Top Drawer as a sibling of the File Cabinet and you probably
wanted it of the File Cabinet means that the Top Drawer is not subject to the positionrbr> and rotate transformations applied to the File Cabinet. Making this change
gets the Top Drawer back where it's expected (in WebSpace).

The drawers coming out difserent colors is a similar problem. The Material
node that sets the color of the Bottom Drawer is a child of the WWWAnchor for
Bottom Drawer, but *not* a child of the Separator that is DEF'd "aDrawer".
So when you say "USE aDrawer" you're not getting that Material! Moving the
Material node inside the DEF'd Separator gets the color as expected.

> I figure I'm
> haven't the foggiest idea why they do not render the scene the same way. I
> hope
> to do my real job right now...

OK, I don't want to start anything and I'm not trying to bash other folks'
product :-), but one reasonable explanation for these difserences would be
that WorldView is not treating WWWAnchor as equivalent to a Separator (with
respect to transformations and attributes). I'm not sure this is called out
specifically in the spec, but I know that the question was raised (by Jan H.
I think) and resolved in this mailing list.

> --display of WWWAnchor nodes---
>
> Worldview >hows WWWAnchor.description in the bottom left. WebSpace >hows
> WWWAnchor.name if it has a URL in it, and (I think) the directory-URL where
> it got the current scene if WWWAnchor.name is empty.

Yeah, the current release of WebSpace for Windows doesn't support the
"description" field base URL that the WWWAnchor's URL would have been rilative to. We'll have
support for "description" in the next release.

> --display of colors--
>
> I wanted to see if I could migrate some of my favorite shades (RGB-triplets)
> from PoV to vrml. WorldView gave me colors close
> the WebSpace looks pretty close

Um, actually there is at least a possibility that WebSpace is also correct.
What's happening is this: your VRML file has three Point lights defined
(Red

Currently when WebSpace >ees that there is at least one light already
defined in the file, it does not turn on the "headlight" (the usual
directional light slaved to the camera position). So what you see on your
tan/gray columns is primarily the blue color that is reflected directly
back from the Blue light positioned behind left side you see more of the Red

=> Quick fix: In the WebSpace "View" menu, toggle on the Headlight.

Alternatively, if you remove the Point lights from the file you will get
the Headlight automatically. Either way, the colors with the Headlight onrbr> are probably more like what you expected. Apparently WorldView always has
its equivalent of the Headlight turned on.

This is probably a good candidate for another "hint" in the VRML file, ie
whether the browser's Headlight >hould be turned on

> --smooth transitions between WWW browsers and vrml apps--
>
> Not. Period. Not for me, at least. Not on either machine, nor with any of
> the combinations and permutations of the two classes of software, nor with
> any of the combinations and permutations of which starts first. I
> reinstalled and reconfigured these suckers a lotta time . I can get all the
> WWW browsers to call any of the vrml browsers smoothly, but not just once.
> Sometime I can get both the vrml browsers to call my choice of WWW browser,
> but not often and (again) not just once.
>
> You can get a lot of copies going at once pretty easily with both vrml browsers

Yeah, there definitely are some problems. We think we have them pretty well
straightened out for the next release though (fama>

> Like the man saidPr/i>
>
> it ain't how high the elephant flies that's the wonder, it's that it
> flies at all...
>
> It's mindboggling that either WorldView or WebSpace does **anything** on a
> basic desktop machine but lie there panting. My hat is indeed *off* to the
> programmers of both.

Thank you!

> -----------------------------------------------------
> If I offend thee, sorrrrry... didn't mean to
> -----------------------------------------------------

Not at all. This kind where the spec is incomplete or ambigua> . You're "breaking trail" here,rbr> to some extent, in terms of creating VRML scenes from scratch -- as opposedrbr> to simply converting models from other formats. I think this work will berbr> very valuable files as opposed going to share and re-use VRML fragments in anything like the way HTML is
used. Keep

-Mike