WebCGM Question

[Lofton Henderson <lofton@rockynet.com>]

CGM Open Members,

I'd like your feedback ASAP -- in case there is still some contradiction or 
ambiguity in the WebCGM Second Release text, we have a chance to fix it 
before WebCGM is published (but this might happen very soon!)

I have a WebCGM question, that has come up during the writing of the test 
suite.  The text of the spec does not help a lot, for example the examples 
of 3.1.3 imply some behaviors, and the bullet list at the end of 3.2.1 
implies different ones.

All WebCGM pictures are required to be 'metric'.  Under what conditions 
should a WebCGM picture be rescaled for display?  Here are some cases to 
illustrate the question:

1. A frameset in a HTML document contains a frame definition:  <frame 
src="myWebCGM.cgm">.  Is the webcgm picture to be initially displayed at 
correct scale, or is it scaled to fit within the frame?  E.g., if the 
drawing is 1m square, is it scaled down for initial display in the frame?

2. A WebCGM instance in a frame contains a hyperlink (linkURI) to another 
WebCGM instance, whole-picture, no object selection or identification (and 
no object behaviors, no view_context, etc).  Same question as #1 -- is the 
initial view of the 2nd WebCGM drawn in the frame/window at correct scale, 
or is it scaled to fit?  (I don't believe that it has an effect, but does 
the answer depend upon _self, _replace, _blank, or 'target' picture behavior?)

3. A WebCGM instance in a frame references an object in another WebCGM 
instance, _self, no object behavior.  Same question -- rescale target 
picture to fit, or draw at correct scale (w/ target object visible)?

4. The only behavior which unambiguously specifies any re-scale is 
view_context -- fit it to display window/frame.  (I think some re-scaling 
potential is strongly implied for highlight_all as well.)

As I mentioned at the start, the examples of 3.1.3 (and figures 3.1 and 
3.2) imply one way:  there is an implication of scale-to-fit, such as in 
the words "The entire picture is displayed..." in the first example.   On 
the other hand, the 3 bullets at the end of 3.2.1.1 seem clear, in its use 
of "move into view" -- preserve proper scale, and move the object into view.

Does the answer differ for whole-picture addressing and object addressing?

So ... what's the answer and what have vendors implemented?

-Lofton.




*******************
Lofton Henderson
1919 Fourteenth St., #604
Boulder, CO   80302

Phone:  303-449-8728
Email:  lofton@rockynet.com
*******************