Re: Liaison with CSS WG to provide a mechanism for expressing the style of document semantics

In trying to make sense of the confusion on this proposal, it is  
probably worth pointing out that semantic-presentation legends are  
already common place in documents. The proposal doesn't add them to  
documents; rather it provides a media independent separation of  
concerns method for authoring them. The HTML5 draft itself makes use  
of one[1]. However, as the proposal makes clear, such semantic- 
presentation legends leave users of other media types out in the cold  
(those using media types that are not typographical for example). They  
also lead authors to resort to poor writing such as: "marked up like  
this" rather than having the UA or the styling author describe "this".

They also require the author to write about presentation and styling  
that the author may have no control over (because the semantics and  
styling are not separated as things stand now). The proposal would  
separate semantics and styling by allowing authors to write documents  
without discussing specific typographic conventions or specific  
semantic-presentation legends.

So perhaps the phrase "typographic conventions" is more familiar to  
those of you trying to understand the proposal, but that phrase is not  
at all a media independent phrase — hence the phrase "semantic- 
presentation legend".

Take care,
Rob


[1]: <http://www.whatwg.org/specs/web-apps/current-work/#typographic>
[2]: <http://esw.w3.org/topic/HTML/SemanticPresentationLegendCSS>


On Jul 31, 2008, at 4:23 PM, Pierre Saslawsky wrote:

>
>> The idea behind this is to allow CSS authors to assign a "legend"  
>> to a particular CSS class. The presentation allows for the "legend"  
>> definition to include text, and the styling of said text in a  
>> variety of formats (on-screen, spoken, etc.). The presentation  
>> itself of the legend is undefined by the draft;
>
> I am impressed with this idea of "Semantic-Presentation Legend" but  
> I see a major shortcoming: as you noted, the presentation itself of  
> the legend is undefined...  I believe the original author should  
> submit a proposal of "SPL-Styling Language" because we certainly  
> don't want to have all the legends styled in a similar fashion. For  
> instance, one should be able to display "This is a joke" in little  
> italic characters, while (to reuse the example of the original  
> document) one may need to show "THE WORDS OF JESUS" all in bold  
> uppercase characters. We may also want to add some generated  
> content, for instance parentheses before and after "(This is a  
> joke)" or a prefix and a suffix to turn "THE WORDS OF JESUS" into  
> "THESE ARE THE WORDS OF JESUS YOUR LORD". The notion of media will  
> take a new importance too because, considering an aural media type,  
> users would be utterly confused or disappointed if these 2 legends  
> were pronounced with the same pitch, volume and elevation.
>
> Thinking about it, as legends become more and more adopted  
> everywhere under each and every element and in so many different  
> styles, we'll probably need to provide the users with some  
> informative texts about the legend presentational attributes. For  
> instance, below the legend "THE WORDS OF JESUS", we could have a  
> little text explaining "This legend appears in bold uppercase  
> letters because the words of Jesus are important" and below the  
> legend "This is a joke", we would have "This legend appears in small  
> italic characters because we don't want you to spend too much time  
> on it". Of course, all this Level-2 legend system would have to be  
> detailed in an even deeper proposal, the "SPL-Styling Language  
> Semantic Legend Module". Although, I'm afraid that once the SPL-SL- 
> SLM takes off, market pressure will push developers to ask for a way  
> to style the Semantic-Presentation Legends presentational attributes  
> informative texts themselves. For instance, below "(This is a  
> joke)", they may want to show "This legend appears between  
> parentheses because SOME PEOPLE JUST DON'T GET IT".
>
> But how would we style these uppercase characters? And how would we  
> explain why they're in uppercase? Hmm... Could someone please make a  
> proposition for a Level-3 Semantic-Presentation Legend system and  
> its associated Styling Language, I'm running out of steam here.
>
> Or could it be that the whole "Semantic-Presentation Legend"  
> proposal itself doesn't make any sense?
>
> Pierre
>
>
>
>

Received on Thursday, 31 July 2008 15:42:58 UTC