ISSUE: clarify about effective region

[Lofton Henderson <lofton@rockynet.com>]

In the process of writing the 'region' ApsAttr test, I found that the 
WebCGM spec could use clarification.

3.2.1.1 (Grobject) talks about what is effectively the "picking region" of 
an object:
-- the 'region' ApsAttr if one is present;
-- else the geometry of the stuff within the Aps;

3.2.2.1 (region) is pretty much consistent, talking about "precedence for 
picking operations".

3.2.2.6 (screentip) talks about "cursor ... over the graphical object", 
without saying exactly what this means.

5.7.10 (WebCGMEvent) talks about "relevant graphical content under the 
mouse", without saying exactly what this means.

Proposed clarification:  all of these indeed refer to the same effective 
region associated with an APS, as defined by 3.2.1.1 and 3.2.2.1:  the 
'region' if one is present, else the geometry within the APS (if any).

Reason.  These all discuss how a mouse operation relates to an APS -- a 
single click initiating associated link execution, a mouseover causing a 
screentip, etc, etc.  They should all behave consistently.

Does anyone disagree?

Btw, the reason this came up when I was working on 'region', is 
interesting.  Some of these DOM tests are hard to verify completely using 
DOM functionality alone.  So I was toying with 'screentip' behavior as 
secondary evidence that the DOM operation had in fact succeeded or failed 
-- i.e., had correctly modified the structure tree or not.  This is 
somewhat risky, as the DOM operation could succeed (as verified for example 
by a 'get...'), but the viewer might not display the result correctly 
(i.e., it could have had a viewing/display problem with the associated 
functionality even at the 1.0 level).  It is not trivial to write tests 
that are "atomic", and isolate the thing to be tested!

-Lofton.