RE: [cgmo-webcgm] ISSUE: grnode

[Lofton Henderson <lofton@rockynet.com>]

At 04:34 PM 3/10/2008 -0400, Weidenbrueck, Dieter wrote:
>  
> I think we need to deal with
> grnodes during traversal.

I suspect that is true.  We just have to be very clear that
visibility in traversal is their *only* DOM presence.

>  
> Consider the following (set to
> CourierNew):
>  
> Picture
>  
> |
> Layer1-Layer2
>  
> |
> grobject1 
> ---------------- grnodeA
>  
> |                          
> |
> grnodeB -
> grobject2         grobject4 -
> grnodeC
>  
> |                                       
> |
> grobject3                               
> grobject5
>  
> If you traverse the tree
> treating the grnodes as normal nodes you have a well
> defined-tree.
> If you disregard the grnode
> nodes, you will never be able to get to grobject4 or grobject5 with
> child/sibling traversal.

Yes, it would be unworkable unless one made some strange special rules,
e.g., for traversal purposes 5 is a child of 4.  Which is a little
too bizarre, I think.

So I won't argue for "completely invisible", unless I think of
some other problem that "only-traversal-visible" raises. 
(And I don't see such now.)

-Lofton.

 

---

From: Lofton Henderson [mailto:lofton@rockynet.com]
Sent: Montag, 10. März 2008 20:02
To: Bezaire, Benoit; cgmo-webcgm@lists.oasis-open.org
Subject: Re: [cgmo-webcgm] ISSUE: grnode

At 12:19 PM 3/10/2008 -0400, Bezaire, Benoit wrote:

If 'grnode' elements are not accessible via DOM calls. Why does WebCGMNode.nodeName lists 'grnode' as a possible value?

Mistake in the document (erratum)?  They happen.

IMO, it is unambiguous that we intended 'grnode' to be *almost* DOM invisible -- it is specified invalid in almost all DOM access.  Did we intend *completely* invisible?  I.e., the question is whether that it participate in the tree traversal results of WebCGMNode [1], even tho' it is invalid in any other DOM calls?

[1] http://www.w3.org/TR/webcgm20/WebCGM20-DOM.html

So there are two normative bits that contradict in Ch.5. 

"...The target APS must not be of type 'grnode'. 'grnode' elements are not accessible via DOM calls"  (while this is in a section describing XCF processing rules, the statement of the second sentence seems more general.)

versus

The DOM accessibility implied in [1].  (Limited to detecting its presence, navigating to its ancestors, siblings, descendents, etc)

Do we agree that that is the scope of the issue?

I don't have a strong feeling about which way it should be clarified.  I have a mild preference for "completely invisible".  But that might have strange implications for navigation (don't know -- haven't thought it through).  If that were the case -- bizarre navigation rules -- I could live with "almost invisible".

-Lofton.