Re: SVG12: feature strings vs "supports"

Hello Björn,

>   http://www.w3.org/TR/2005/WD-SVGMobile12-20050413/feature.html seems
> to be quite inconsistent with respect to feature strings for attributes
> and properties.

We have reviewed the features appendix, but do not find all of the
problems that you note.

"In some cases the feature strings map directly to SVG modules, in
others they represent some functionality of the User Agent (that it is a
dynamic viewer for example)."

>  For example, image-rendering, shape-rendering, and text-
> rendering have associated feature strings but color-rendering does not.

Thank you, color-rendering was missing from the Paint Attribute module
and thus was not covered by the corresponding feature string; we have
corrected this.

> More generally, it is not clear when requiredFeatures for a given string
> would return true; that some attributes have feature strings seems to
> imply that it might be possible for an implementation to return true for
> http://www.w3.org/Graphics/SVG/feature/1.2/#Shape even though it does
> not support http://www.w3.org/Graphics/SVG/feature/1.2/#PaintAttribute
> it is however not clear whether this is the case.

"indicates that the viewer can process and render successfully all of the
corresponding language features" seems fairly clear. It would be hard to
render shapes without painting them. We don't claim that it is possible
to produce an implementation with an arbitrary sequence of true and
false for its feature strings; some depend on others.

> For http://www.w3.org/Graphics/SVG/feature/1.2/#SVG-all it is not clear
> whether http://www.w3.org/Graphics/SVG/feature/1.2/#SVG-static-DOM must
> return true aswell, the draft only notes that all the features that
> correspond to the latter must be supported;

which would therefore cause it to return true, no?

>  the definition for
> http://www.w3.org/Graphics/SVG/feature/1.2/#SVG notes "At least one of
> the following" must be supported, clearly that would be satisfied if an
> implementation returns true for #SVG-all but not for #SVG-static-DOM.

"At least one of" a set of four seems fairly clear to us. Its false if
zero are supported and true if one or more is supported. We also note
that this is a pretty general "SVG is supported" feature string and
recommend using more specific tests:

"it is recommended that more specific feature strings be used"


> http://www.w3.org/Graphics/SVG/feature/1.2/#SVG-static-DOM for example
> corresponds to the <script> and <handler> elements. It is not clear when
> these are considered "supported", does it mean that the user agent can
> successfully ignore the elements,

no

>  or that it supports one language
> binding for use with these elements and all SVG uDOM features?

yes. Otherwise it could not "process and render successfully all of the
corresponding language features".

> Please change the draft such that all of this is explicitly defined.

We have fixed the specific deficiency that you noted, and do not see any
imprecision for the more general points you make.

Please let us know shortly if this response does not satisfy your
comment.

-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Thursday, 10 November 2005 22:10:30 UTC