Re: [SVGMobile12] 6.2: "SVG Tiny 1.2 does not support CSS selectors"

On Thursday, May 19, 2005, 3:47:02 PM, Ian wrote:


IH> SVGT1.2 section 6.2 says:

IH> # SVG Tiny 1.2 does not support CSS selectors applied to SVG content.

IH> I'm not sure what this statement is trying to say. Please either remove 
IH> this statement or change it to make sense.

The statement was added at the request of other commentors, who found
earlier statements that SVG Tiny "did not support CSS "to be
insufficiently precise. For example, some attributes have the same
syntax as the relevant CSS property. The part of CSS that Tiny does not
use is the CSS selector syntax. The only way to set a property on an
element is via a presentation attribute, or by animating the property.

IH> A literal reading of the statement implies that the specification itself
IH> is not an implementation that supports CSS selectors, but that is a 
IH> trivially true statement since the spec isn't an implementation, but a 
IH> spec.

We agree that a spec is not an implementation.

IH> As far as I can see these are the various things that statement 
IH> could be trying to say:

IH>  1. "User agents are not required to implement the CSS cascade": That much
IH>     is obvious, since this is the SVG spec, and not the CSS spec.

Its not the XML spec, either,and yet all SVG Tiny implementations are
required to implement an XML parser. You are correct that SVG user
agents are not required to implement CSS selectors.

IH>  2. "User agents must not implement the CSS cascade": I doubt this is what
IH>     is intended, as it would be inappropriate for one spec to put 
IH>     restrictions on what other specs may be implemented.

You are correct that the spec does not say this.

IH>  3. "Authors must not rely on the CSS cascade to style documents that
IH>     are intended to be used with user agents that only support SVG Tiny 
IH>     1.2": That seems redundant, since the whole point of the CSS cascade
IH>     is that the styling is optional.

You are correct that authors must not rely on it; this is in fact worth
saying.

Styling may be optional with textual documents, butthe utility of a
graphic where everything was drawn black on black would be minimal.
Extrapolating from experience with HTML to all other XML formats is not
necessarily wise.

IH> Since none of those three possibilities make any sense to me, I must 
IH> assume that there is a fourth meaning that eludes me.

We are saddened that the simple statement that CSS selectors are not
required makes no sense to you. However, your proposed text

  "Authors must not rely on the CSS cascade to style documents that are
  intended to be used with user agents that only support SVG Tiny 1.2"

while apparently not making sense to you, does make sense to us and we
propose to add it to the specification.

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

Received on Sunday, 30 October 2005 03:34:29 UTC