I prefer to code with markup to the strictest of the rules.
XML and XHTML 1.1 are my markup languages of choice because of their well-formedness constraints and strict rules.
I understand HTML5 is the new kid on the block, but HTML5 (to me) is nothing more than HTML 4.01 with a boob job. Repurposing the B and I tags using long-winded phrases from academia is as fake as Joan Rivers’ looks.
ARIA 1.0 has been in development for many years. ARIA supplies attributes to help assistive technologies (AT) better understand the web author’s intent in markup.
ARIA is supposed to make visible to AT those parts of a web document previously invisible.
I would say read the ARIA 1.0 recommendation to learn more about the attributes and their purpose from the people who know, but this is the problem. This is the reason for my frustration.
ARIA 1.0 is BULLSHIT! Plain and simple.
After all these years, the ARIA 1.0 specification reached recommendation status. Inside the spec, the recommendation is still referred to as a draft (take a quick look at chapter 1).
Chapter 10 mentions an XHTML + ARIA DTD and states that documents can be validated using that DTD. Nowhere else in the spec does it say otherwise.
The W3C validators do not recognize XHTML + ARIA. Let me put this in perspective.
In the Web Content Accessibility Guidelines (WCAG) 2.0, under Success Criteria 4.1.1, two of the sufficient techniques provided by the W3C are G134 (Validating web pages) and G192 (Fully conforming to specifications).
Simply put, if I were to add ARIA attributes to an XHTML document that previously met single A conformance, it would now fail miserably.
In other words, using a recommendation from the W3C with one of its own host languages will cause my XHTML 1.1 documents to fail WCAG 2.0 single A conformance.
Why? Because XHTML 1.1 + ARIA 1.0 will not validate and, based on some feedback that supposedly is from a reliable source, it will never be made to validate.
On the surface, this is no big deal.
The problem is that no word was put out from the W3C, especially to all those using XHTML Family of markup, that ARIA 1.0 will not apply to the host language (XHTML).
This is like the store manager looking at a long line of customers waiting for 10 days for the new video console and finally decides to open the store and say, “We will sell only to those driving a Ferrari.”
“Hey! Did you not see us waiting for this?!”
Specifications are supposed to specify, in detail, the technology. Period. No inferences need to be made by the reader. The detail is all in the specification and, though boring, explicit in nature.
A specification should not waste anyone’s time mentioning things that are extraneous or irrelevant. Specify and be specific!
The esoteric nature of the recommendations that come out of the W3C is bad enough. As much as their own WCAG 2.0 asks developers to acknowledge those with cognitive shortcomings, recommendations seems to fail in this miserably.
Still, with the aristocracy of the W3C editing the specs, I sometimes can hear their voices saying, “These recommendations are really not for the hoi polloi, but provided to you as a generous offer from on high.”
If the W3C wants the web to move forward, then making sure those in the trenches bringing folks to the web with their great designs and applications understand both the intent and practice of their (W3C) technologies.
A specification that cannot be relied on is not a specification, but a fairytale.
I prefer XML 1.1, XHTML 1.1, RDF 1.1, and RDFa 1.1. I am still reluctant to actually use RDF or RDFa (I don’t really want to make it easier for the NSA to find information).
After waiting all these years for ARIA 1.0 and realizing its unfulfilled promise, I am left with only one question:
What’s the point?