Default Object Behavior

[Lofton Henderson <lofton@rockynet.com>]

CGM Open Members --

We have to come to a resolution on this issue.  I think a conference will 
be required.

Last week we had some discussion about clarifying some default object 
behaviors.  The scope opened up a bit, and Dieter has made a suggestion for 
a change to the WebCGM 1.0 specification for Second Release (see his email 
w/ attached tables).  In particular, there are cases where the behavior is 
currently clearly specified, where Dieter requests a change:  basically 
from "full-picture view, highlight object", to "zoom-to-object, highlight 
object".

I implied last week that my own position was -- don't care, but procedural 
objections.  Upon thinking about it from a user perspective, I have decided 
that I do have a preference, and that is:  per current spec (full-picture 
view & highlight).  I have talked with at least one other user who feels 
the same (apparently strongly).

My informal survey of implementors show that all others do it per the 
current spec.

So this needs to be talked out, and we need to have a clear majority 
supporting change.  Furthermore, I suspect that any single resolution will 
leave someone unhappy.  Therefore...

I have been pondering a compromise position:

1. First (and somewhat independently, but related) the spec needs to say 
something like, "In interaction-capable environments, WebCGM viewers shall 
be capable of zoom and pan operations, and shall offer zoom and pan 
controls to the user."  This is implicit, I think, but we never say 
it.  The words I chose are similar to what SVG says.

2. Now the compromise.  We could leave the spec as is (except to clarify as 
needed) and define it as default behavior of conforming viewers.  But we 
could add something like:  "Viewers shall (should?) have an operating mode 
or option whereby [...reverse the default to "zoom-to-object" on the 
disputed cases...]"  I didn't fill in the words of the [...] portion, but 
you get the idea.  Viewers would have to be capable of doing it both ways, 
with current 1.0 First Release spec as the normal default.

Does this look like a useful approach?  (It is an odd specification to put 
in, but perhaps we could add a non-normative note indicating that it is for 
the purpose of "accommodating certain legacy implementations and content".)

Who wants to participate in a conference and when (it needs to be Thursday 
or later)?

-Lofton.

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

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