Re: [cgmo-webcgm] ISSUE: relative URIs in 'linkuri'

[Benoit Bezaire <benoit@itedo.com>]

Maybe this was discussed on the telecon; but here are my thoughts
anyway...

I don't understand what the problem is; or what is meant by 'linkuri'.
'linkuri' the WebCGM 1.0/2.0 APS attribute; or <linkuri> element of
the XCF?

If we are talking about the WebCGM 1.0/2.0 APS attribute, then this is
easy, relative URIs should behave in 2.0 liked they behaved in WebCGM
1.0. I'm sure you'll all agree.

I suspect Dieter was talking about the <linkuri> element of the XCF.
Please don't forget that applying an XCF onto a WebCGM document is
'replacing' the existing WebCGM APS attribute values with the new ones
found in the XCF. When an implementation sees:

<grobject apsid="one">
  <linkuri uri="#myObject" />
<grobject>

The string '#myObject' is set within the 'one' APS, like so:
BEGAPS 'one' 'grobject' STLIST;
  APSATTR 'linkuri' "14 3 '#myObject' '' ''";
  BEGAPSBODY;
  ...
ENDAPS;

And the hyperlinking behavior of such a file should be well defined
(as per WebCGM 1.0).

Regards,

-- 
 Benoit   mailto:benoit@itedo.com

 
Monday, September 19, 2005, 6:49:49 PM, Lofton wrote:

LH> Ref: 
LH> http://lists.oasis-open.org/archives/cgmo-webcgm/200509/msg00093.html
LH> Comments:  20
LH> ==========

LH> ISSUE:  how are relative URIs in 'linkuri' resolved?

LH> DISCUSSION:  Dieter writes,
LH> "something should be added to say that xcfurls can be relative but can also
LH> be absolute. Also, we need to make a comment about relative URLs found in
LH> linkURIs then, the processing there is different, if I remember 
LH> correctly.There it is relative to the container document I think."

LH> The first sentence is strictly editorial.  The second is not so 
LH> straightforward.  A 'linkuri' could occur:
LH> -- in a WebCGM instance;
LH> -- in DOM script;
LH> -- in XCF instance.

LH> As I understand it, the container document would be the WebCGM, a HTML/JS
LH> file, the XML file of the XCF.  I don't think we want the relative URI of a
LH> 'linkuri' to resolve to the container document.  For example, load this CGM

LH> Container:  http://www.example.org/webcgm/myCGM.cgm
LH> linkuri:  #another-obj-in-this-pic

LH> Then we edit the DOM of that already-loaded WebCGM, with this DOM script,
LH> to change that linkuri...

LH> Container:  http://www.example.org/domScripts/myDOMtest.html
LH> linkuri:  #a-third-obj-in-this-pic

LH> I don't think we want the linkuri to all of a sudden be relative to
LH> myDOMtest.html.  Or am I misunderstanding the meaning of "container
LH> document"?  I think we want something like ... relative URI in 'linkuri'
LH> resolves relative to the WebCGM "that contains it".  I quoted the latter
LH> because I'm not quite sure the right way to express it.  But ... if the
LH> 'linkuri' is in a WebCGM, no problem.  If it is in DOM script or XCF ...
LH> well, the 'linkuri' in a sense is really contained in the WebCGM, and those
LH> (DOM script / XCF) are just editing directives that are being applied, to
LH> change the 'linkuri' value in the WebCGM.

LH> RECOMMENDATION:  relative URI in 'linkuri' resolves relative to the WebCGM
LH> "that contains it".

LH> Discussion?  (Suggestions for the best way to say it?)

LH> Regards,
LH> -Lofton.