Re: Reconsider SVG 1.2

On Thursday, November 18, 2004, 8:59:29 PM, Ian wrote:


IH> On Thu, 18 Nov 2004, Chris Lilley wrote:
>> 
>> most worrying of all, apparently the CSS WG (or some vocal spokespeople
>> for it) believe that CSS should not be used with random XML at all - 
>> only HTML should be used.

IH> To clarify: the assertion to which you refer is that CSS is an optional
IH> layer,

Yes

IH>  and that the markup sent over the wire should be in a language that
IH> is natively recognised by the user's agent.

Why should the user agent have anything to do with it? Its whether its
recognised by the user that is impostant. if, for example, a geologist
wants to get some material in RockML than, frankly, if its a
textual-document-like markup language where the CSS box model would do a
reasonable stab at laying it out, then sending RockML plus a CSS
stylesheet smmes a way better thing to do than, say, converting it to
HTML (thus loosing all of the semantics) before transmission.

IH> This isn't a CSS matter -- it's basic accessibility, it applies whether
IH> CSS is involved or not. WCAG guidelines 9 and 11 are based on this 
IH> concept:

IH>    http://www.w3.org/TR/WAI-WEBCONTENT/#gl-device-independence
IH>    http://www.w3.org/TR/WAI-WEBCONTENT/#gl-use-w3c

Having read those and determined that XML is a W3C technology and that
any imagemaps should be client side and so forth, i don't se how RockML
would be a problem for the geologist there.

IH> So it's not that "CSS should not be used with random XML at all", it's
IH> simply that you shouldn't use random XML at all in the first place.

By random I meant, something that is random (ie unrecognized) XML from
the pov of the user agent. its not random to the content creator or the
person using it, of course.

IH> IMHO, CSS works well with SVG, for example it allows SVG links to be a
IH> different colour based on whether the link is visited or not, and allows
IH> applications to give hover feedback easily,

Yes, the dynamic pseudoclasses certainly have value in CSS. Although,
they would have equal value in an XML syntax of course.

IH>  with alternate stylesheets it
IH> allows different looks to be easily applied to the same basic shape, and
IH> so forth. Much the same advantages that HTML gets from CSS.

Well, I agree that the advantages that SVG gets from CSS also apply to
HTML. The reverse is not true. In HTML or other document-like techdoc
type languages, if you have a thousand paragraphs, the chances are they
are all styled the same, and if not, within a section they are all
styled the same. For an SVG document with a thousand paths, the chance
that any two share the same fill color is not that great.

IH>  Of course CSS
IH> isn't _required_ for either; it is, by design, an optional layer.

Right. Although that implies that keyboard shortcuts and tab order and
so forth should not be defined in CSS, per
http://www.w3.org/TR/WAI-WEBCONTENT/#gl-device-independence.

-- 
 Chris Lilley                    mailto:chris@w3.org
 Chair, W3C SVG Working Group
 Member, W3C Technical Architecture Group

Received on Thursday, 18 November 2004 21:37:46 UTC