W3C home > Mailing lists > Public > www-svg@w3.org > September 2008

Re: [1.2T-LC] foreignObject content and referenced resource (ISSUE-2062)

From: Doug Schepers <schepers@w3.org>
Date: Sat, 20 Sep 2008 12:32:27 -0400
Message-ID: <48D5259B.9030903@w3.org>
To: www-svg@w3.org
CC: "Dr. Olaf Hoffmann" <Dr.O.Hoffmann@gmx.de>

Hi, Dr. Olaf-

Thanks for your insightful comment.  I've raised ISSUE-2062 in our
Tracker, and the SVG WG will discuss it and get back to you as soon as
possible.

Thanks-
-Doug, on behalf of the SVG WG


Dr. Olaf Hoffmann wrote (on 9/20/08 9:34 AM):
> Hello SVG WG,
> 
> doing some experiments with foreignObject I discovered, that there is some
> information missing in the definition or description of foreignObject.
> This should be clarified.
> 
> 
> 1. Compared to SVG1.1 the use of xlink:href is new, but in the attribute 
> definitions is not mentioned, what should happen with the referenced 
> resource (what can be almost everything according to 14.1.4 Reference 
> restrictions), if a user agent decides to interprete foreignObject 
> somehow.
> 
> For example for a referenced XHTML document (as well as for
> embedded XHTML) Opera 9.5 decides to display the top left corner 
> fragment of the referenced document into the area described by 
> x, y, width, height attributes following the painters model (earlier 
> other elements below, later elements on top).
> Other parts of the referenced document beyond this area are not
> accessible.
> 
> My guess is too, that it is at least intended, that the referenced 
> document or a fragment of it is somehow displayed inside the
> given area (following the painters model), if displayed at all.
> But then this should be mentioned in the attribute definition 
> as for image, use, animate, set etc or in the general 
> description of the element, as for example for audio and video.
> 
> Another interpretation could be, that the xlink:href references
> an alternative view, a description (xhtml:longdesc), something 
> related to the foreignObject content. Or the content of 
> foreignObject could be an alternative fallback for the 
> referenced document similar to the behaviour of xhtml:object. 
> 
> (Side note: the element animation has a similar problem, it
> is not mentioned, what the purpose of xlink:href is, however
> this can be guessed, but I would like to suggest to replace
> "The referenced SVG document represents a separate document 
> which generates its own parse tree and document object model."
> with
> "The SVG document referenced with the xlink:href attribute 
> represents a separate document which generates its own 
> parse tree and document object model."
> to cover this problem for the animation element, a similar
> description could solve parts of problem for foreignObject as
> well).
> 
> 
> 2. According to the SVG1.1 usage and the example
> 23_02.svg it is possible too to put the foreign content
> directly inside the foreignObject.
> If additionally an external resource is available with
> xlink:href, what happens? What is the intended 
> relation between the content and the document
> referenced by xlink:href? 
> Is it an arbitrary choice of the viewer, what will be
> displayed if the viewer decides to display something
> and we assume that both contents are displayable for
> the viewer somehow?
> a) Is the foreignObject content displayed only?
> b) Is the referenced document displayed only?
> c) are both content and referenced document displayed?
> c1) is the content on top (because of the painters model and
> the xlink:href is slightly earlier in the source code than the 
> content?)
> c2) is the referenced document on top?
> d) should the viewer provide an interface to select, what
> to display?
> 
> 3. Obviously different from SVG1.1 in SVGT1.2 the
> foreignObject currently does not create its own
> viewport
> http://www.w3.org/TR/SVGMobile12/coords.html#ElementsThatEstablishViewports
> 
> How is the foreign content aligned relative to the
> user coordinate system? Is there a chance for the author
> to influence, which part of the referenced content is
> displayed, if a viewer like Opera 9.5 displays only a fragment
> or is this the problem of the foreign language (for example
> XHTML, typically creating scrollbars or other mechanisms 
> automatically to get access to the content of the complete 
> document)?
> Is the content outside the x, y, width, height area always
> clipped or is it on the viewer to decide this or on the author
> to avoid this?
> 
Received on Saturday, 20 September 2008 16:33:01 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:40 GMT