Re: AuthConfReq: Presentational Markup

On Mar 27, 2010, at 5:17 AM, Sam Ruby wrote:

> The reasons given for disallowing presentational markup given are:
> - The use of presentational elements leads to poorer accessibility
> - Higher cost of maintenance
> - Higher document sizes
>
> To explore this, I offer this nearly perfect specimen of markup:
>
>  http://diveintomark.org/archives/2007/06/30/irony
>
> And draw attention to two parts:
>
>  <b style="background:transparent;color:red">1984</b>
>  <strike>the</strike>
>
> The former conforms to the author conformance requirements present  
> in the document.  How does this lead to greater accessibility than  
> the alternative?  How does it reduce maintenance costs?  How does it  
> reduce document sizes?
>
> The latter does not conform to the author conformance requirements  
> present in the document.  How is this less accessible than the  
> alternative?  How does it increase maintenance costs?  How does it  
> increase document sizes?

It should be noted that the rationale for author conformance  
requirements explicitly calls out the style attribute as a piece of  
presentational markup that is allowed notwithstanding the general  
reasons for the ban:

<http://dev.w3.org/html5/spec/Overview.html#presentational-markup>

"The only remaining presentational markup features in HTML are the  
style attribute and the style element. Use of the style attribute is  
somewhat discouraged in production environments, but it can be useful  
for rapid prototyping (where its rules can be directly moved into a  
separate style sheet later) and for providing specific styles in  
unusual cases where a separate style sheet would be inconvenient.  
Similarly, the style element can be useful in syndication or for page- 
specific styles, but in general an external style sheet is likely to  
be more convenient when the styles apply to multiple pages."

So you could argue that this exception is not well justified (and  
style should be banned too), or that the rationale for any  
presentational markup to be banned is not well justified, or that the  
style attribute and stye element are the wrong place to draw the line.  
However, I am not sure it is instructive to compare these two pieces  
of markup against the criteria for presentational markup in general.  
The spec explicitly acknowledges that the style attribute may suffer  
all the problems of presentational markup, but gives specific reasons  
for granting it an exception.

Second, I am not sure examining an individual case in isolation is the  
best way to investigate the validity of this item of rationale. The  
rationale in Section 1.9.1 does not say that every single use of  
presentational markup leads to poor accessibility. In fact, it  
explicitly acknowledges that in some cases, use of presentational  
markup may lead to acceptable accessibility outcomes. What it argues  
is that use of presentational markup, particularly habitual use to the  
exclusion of semantically appropriate elements and CSS, *tends* to  
result in worse accessibility outcomes. The argument, as I understand  
it, is that free use of presentational markup creates negative  
externalities, therefore it should be disallowed except in cases where  
there is a good reason to allow a particular construct.

Since we've had a few religion-themed analogies on this thread  
already, perhaps the one to cite here is "building a wall around the  
Torah". This term refers to the Jewish religious rulemaking tradition  
where the original proscription is quite narrow, but Rabbis have over  
the years created a broader rule that is perhaps more restrictive than  
necessary, but where it is much easier to tell if you are actually  
following it, and harder to violate it by itself. For example, the  
biblical commandment, "thou dost not boil a kid in its mother's milk"  
is expanded to a rule where no meat may be eaten with any dairy  
product, or even from the same plate.

Banning <font> in general, rather than, say, only when used in a way  
that actually harms accessibility, is analogous to this reasoning. By  
having the blanket ban, we avoid the presumed negative externality,  
without having to closely inquire about the particular circumstances  
of each use. The latter requires too much judgment for a conformance  
checker.


Note: I am not sure if I personally accept the argument against  
presentational markup in full. I am still mulling it over. In  
particular, I think there may be presentational constructs where the  
accessibility argument almost never applies, and where benefits to the  
author are highly dependent on specific circumstances, to the point  
that a blanket ban does not help authors on the whole. Also, I am not  
sure we have necessarily identified all cases where the potential  
benefit to using a particular construct exceeds the negative  
externality. However, I think it is important to fairly present the  
argument made in the spec, and to address it on its own terms.


Regards,
Maciej