Harry,
Thanks for your thoughtful contribution to the discussion about default
object behavior. I'd like to briefly summarize how I see the situation.
I hope others (esp the implementors) will speak up also.
We are discussing, specifically, the behavior of viewers in the situation:
* a link is executed to an object (APS) in a WebCGM picture;
* there is no object behavior associated with the link;
* there is no 'viewcontext' APS attribute on the object.
* there is no 'region APS attribute on the object (this is somewhat
superfluous).
Dieter's suggested that the behavior specified in WebCGM 1.0 (First
Release, 1/99) be changed for WebCGM 1.0 Second Release. The behavior
specified is: highlight the object in a full-picture, unzoomed
view. Dieter suggests: zoom the view (and "flash").
Summary of my position:
1. The suggestion does not fix a defect (at least not a serious one that
prevents WebCGM from working);
2. It changes a behavior which has been discussed before (1998 and again
2000) and deliberately put into the standard, and for which there are
reasonable supporting arguments. I.e., it is one preference for how things
should work, but not the only reasonable option.
3. We are working on 2nd Release of WebCGM 1.0 -- **1.0**, same functional
version of standard. We are progressing it through W3C under expedited
rules that relieve us of too much attention, PR review, etc. I get nervous
every time we discuss something that looks like a functional change.
4. I know about the behavior of two implementations -- one does it per the
standard, the other does it per the suggested change. (I'm surveying to
see how the other implementations do it).
5. It is legal and valid for a viewer to offer an optional mode,
"always-zoom-target-objects", as long as it supports the standard as well
(and I'd say the standard would need to be the default mode).
Summary: "If it's not broken, don't fix it". (And, "let's be done with
Second Release").
Having said that, I would not resist further if the WebCGM community were
unanimous for the change, and/or if all (most) implementations unanimously
do it per the suggested change.
Looking at some of your specific comments...
At 08:19 AM 5/30/01 -0400, Whittaker Harry W CRBE wrote:
>[...]
>It appears everyone has been working very hard on CGM issues.
>I have been keeping up to date on the recent flurry of e-mail
>messages regarding the specification. So Lofton I take it that
>because of procedural changes we are currently tabling some of
>Dieter's recent suggestion to the specification? I understand
>all to well what it can do to an author or document when people
>provide comments and suggestions past the cut off date. But I
>have a serious question to the CGM community and ALL the implementers
>in our group.
Again, by "procedural" I'm really summing up my above points 1-5: I don't
see that anything is broken, it is a considered and deliberate spec in the
standard, and even if the suggested changes look appealing, we should be
careful about re-opening and changing this -- only if there's a strong
majority in favor (unanimous?).
>Taking into account the default behaviors using
>view_context, and object default behaviors, and high light, what
>are the products who have viewers and have viewers integrated doing?
This is the most important point. I'm surveying now.
>I mean if most of these products have these solutions already coded
>and as a community we agree that we like these behaviors,
I agree. If the implementations all do it contra what the WebCGM 1.0 First
Release says, then it would be sensible to find a way to slip in the change.
> we might
>think about delaying the new release a week or two or even a month.
>(sorry Lofton)
The co-editors (Dave and me) would survive, even though our original target
date was 1/22/01!
>I am not suggesting a reversal of any previous group
>decision.
I believe that it was a deliberate (but probably not unanimous) decision --
but perhaps my memory is not accurate. I do know that it has been
discussed and debated, both in the original writing of WebCGM and last year
during "Second Release" discussions.
-Lofton.
*******************
Lofton Henderson
1919 Fourteenth St., #604
Boulder, CO 80302
Phone: 303-449-8728
Email:
*******************
