} On Nov 13, 12:03pm, Antheopohedron wrote:
} > Subject: VRML 1.1 proposed changes
} > ...
} >
} > 2) The Texture2 imagemapping I was talking about before. This hit the
} > mailing list, was supported briefly, and vanished like a drop in the
} > bucket. I thought I had made a good case for it. Again, are there any
} > objections or reasons why it is not feasible? Please seriously
} consider
} > it.
}
} The better place for "imagemapping" to be done is in the WWWAnchor node,
} not the Texture2 node. There was a proposal for this a while back, but I
} don't think it is currently under consideration. It would simply add a
} new enumeration (such as TEXTURE) to the map field. This would send the
} texture coordinates rather than the object space point clicked on. This
} would be sent in normalized coordinates (0..1) but these are easily
} converted to image space to duplicate the HTML imagemap capability.
The problem with returning normalized coordinates and having the server
(CGI script) translate them is twofold:
1) You cannot use the existing imagemap CGI, which works quite nicely. It
is fairly reasonable to expect someone to use an imagemapped picture
both in an HTML document and as a texture in a VRML world. An authro
would want to be able to use the same imagemap file and the same
imagemap CGI script, and they SHOULD be able to.
2) Even if we accept that there should be a new imagemap CGI script to
handle this (or perhaps just a modification to the old one to act
differently if it receives different arguments... more on that below),
there is the problem of making the height and width of the texturemap
available to the CGI script. Will your browser send
xnorm,ynorm,width,height? Will you have to add some kludge like giving
the width and height of the image in the first line of the imagemap
file? Or in the name of the imagemap file (which is the kludge I am
using for this QTVR thing I mentioned before... still hasn't been
tested)?
In the interest of sticking to existing mechanisms, PLEASE make it
identical to the HTML imagemap system. The browser can certainly afford two
floating point multiplies and two roundings.
} --
} chris marrin
--Greg