To Hell with WCAG 2 – A Listing Aside

by

The Net Content material Accessibility Pointers 1.0 have been revealed in 1999 and rapidly grew outdated. The proposed new WCAG 2.0 is the results of 5 lengthy years’ work by a Net Accessibility Initiative (WAI) committee that by no means fairly received its act collectively. In an effort to be all issues to all net content material, the basics of WCAG 2 are practically unattainable for a working standards-compliant developer to know. WCAG 2 backtracks on fundamentals of accountable net growth which might be effectively accepted by standardistas. WCAG 2 isn’t sufficient of an enchancment and was not well worth the wait.

Article Continues Under

Put together for disappointment#section2

In the event you’re a standards-compliant net developer, you already learn about net accessibility and are conversant in the one worldwide normal on that subject, the Net Content material Accessibility Pointers. WCAG 1 simply celebrated its seventh birthday and is closing in on the top of its life. WCAG 1 badly wants revision.

On 27 April 2006, WAI revealed the primary instalment of the interminable sequence of paperwork required for the revision, WCAG 2.0, to change into a normal.

In the event you have been hoping for a wholesale enchancment, you’re going to be dissatisfied. Numerous unfastened ends have been tidied up, and lots of low-priority tips are actually fairly stable. The issue right here is that standardistas already knew what to do to cowl the identical territory as these low-priority tips. The place WCAG 2 breaks down is within the huge stuff. Curiously, although, and maybe attributable to meticulous enhancing over time, the massive stuff is effectively camouflaged and, to an uninformed reader, WCAG 2 appears cheap. It isn’t, and also you as a working standards-compliant developer are going to seek out it subsequent to unattainable to implement WCAG 2.

The place to seek out the paperwork#section3

Within the nice custom of the W3C, the precise WCAG 2 paperwork are complicated and exhausting to find. (I’ll additionally offer you pagecounts, as printed to U.S. letter–sized PDF from Safari with unchanged defaults, in addition to wordcounts with out markup.) I printed and skim all three of those paperwork for this text.

  • Net Content material Accessibility Pointers 2.0 is the precise root doc and is the one one that’s “normative,” i.e., a normal. It’s described, in W3C parlance, as a Final Name Working Draft. (72 pages, 20,800 phrases)
  • Understanding WCAG 2.0 is a doc that purports to elucidate WCAG 2. (165 pages, 51,000 phrases)
  • Strategies for WCAG 2.0 gives “common” methods. (221 pages, 88,000 phrases)

When put next towards typical web page dimensions in books, the three WCAG 2 paperwork, at 450 pages, exceed the scale of every of the books revealed on the subject of WCAG 1, together with mine. Moreover, in keeping with many weblog experiences (Snook, Clagnut, Sitepoint), Shawn Lawton Henry of the WAI Schooling & Outreach Working Group cautioned attendees at her South by Southwest 2006 presentation to learn solely the Understanding doc, not the precise spec. Because the Understanding doc is greater than double the scale of what it purports to elucidate, this itself might point out an issue with WCAG 2.

There’s a separate doc, not up to date since November 2005, overlaying HTML methods. It isn’t included on this article. Additionally, “tips” in WCAG 1 are actually referred to as “success standards” in WCAG 2, a change in nomenclature I’ll ignore.

You don’t have a number of time to remark#section4

After engaged on WCAG 2 for 5 years, WAI gave the complete trade and all events, together with individuals with disabilities, a whopping 34 days to touch upon WCAG 2 (till 31 Could 2006). Whereas that’s in extra of the prompt three-week minimal, it isn’t lengthy sufficient. The Working Group, furthermore, would love you to fill out a type, probably utilizing Excel, for every concern you disagree with.

I counsel you to easily ship mail to [email protected] and skim the archives of that mailing record (the place it’s unattainable to inform precisely who submitted what remark by way of the WAI type). There’s a prolonged omnibus record of feedback acquired by way of the WAI type. I additionally advise individuals to petition for no less than one other month’s commenting time, quoting W3C course of again to them (viz., remark intervals “might last more if the technical report is advanced or has important exterior dependencies”).

And now a phrase about course of, which you’ve gotten have to understand with the intention to perceive the outcome. The Net Content material Accessibility Pointers Working Group is the worst committee, group, firm, or group I’ve ever labored with. A number of of my buddies and I have been variously ignored; threatened with ejection from the group or truly ejected; and actively harassed. The method is stacked in favour of multinationals with expense accounts who can afford to speak on the cellphone for 2 hours every week and jet to world capitals for conferences.

The WCAG growth course of is inaccessible to anybody who doesn’t converse English. Extra importantly, it’s inaccessible to some individuals with disabilities, notably anybody with a studying incapacity (who should wade by means of ill-written requirements paperwork and e-mails—there’s already been a criticism) and anybody who’s deaf (who should hearken to convention calls). Nearly no one with a studying incapacity or listening to impairment contributes to the method—as a result of, in sensible phrases, they’ll’t.

What WAI is supposed to be doing is enhancing the online for individuals with disabilities. One thing’s improper if many members work in a local weather of worry, as they inform me they do. I by no means hear of comparable complaints from WAI’s different teams. WCAG Working Group is a rogue ingredient inside the W3C, one which chair Tim Berners-Lee should urgently carry to heel.

The method is damaged, so let’s not be stunned that the results of that course of is damaged, too.

Much less of a travesty, however nonetheless a failure#section6

In the event you ever put aside two hours of your life to learn a earlier “draft” of WCAG 2, you have been most likely baffled and/or infuriated. The Working Group has been efficient at enhancing minor tips and has excelled at making the entire doc appear eminently cheap. They’ve succeeded spectacularly at burying the lede—hiding the nub of the rules deep inside the doc. They’ve accomplished a wonderful job at making WCAG 2 appear like it is going to truly work. It received’t.

Based mostly on the three paperwork I learn, taking into tài khoản each required and prompt practices, let me clarify what WCAG actually says:

  1. Precisely what a “web page” is, not to mention a “web site,” will probably be a matter of dispute.
  2. A future web site that complies with WCAG 2 received’t want legitimate HTML—in any respect, ever. (Extra on that later.) You’ll, nonetheless, should test the DOM outputs of your web site in a number of browsers and show they’re an identical.
  3. You’ll be able to nonetheless use tables for format. (And never simply a desk—desks for format, plural.)
  4. Your web page, or any a part of it, might blink for as much as three seconds. Components of it could not, nonetheless, “flash.”
  5. You’ll have the ability to outline whole applied sciences as a “baseline,” which means anybody with out that know-how has little, if any, recourse to complain that your web site is inaccessible to them.
  6. You’ll have the ability to outline whole directories of your web site as off-limits to accessibility (together with, in WCAG 2’s personal instance, all of your freestanding movies).
  7. In the event you want to declare WCAG 2 compliance, it’s essential to publish a guidelines of declarations extra paying homage to a compelled confession than any of the accessibility insurance policies usually discovered in the present day.
  8. Not that anyone ever made them accessible, however in the event you put up movies on-line, you now not have to supply audio descriptions for the blind on the lowest “conformance” stage. And solely prerecorded movies require captions at that stage.
  9. Your podcasts might should be remixed in order that dialogue is 20 decibels louder than prolonged background noise. (You don’t should caption or transcribe them, since they aren’t “multimedia” anymore. Nonetheless, slideshows are actually formally deemed to be “video,” which can come as a shock to Flickr customers.)
  10. You’ll be able to put a number of hundred navigation hyperlinks on a single web page and do nothing extra, however if in case you have two pages collectively which have three navigation hyperlinks every, it’s essential to present a solution to skip navigation.
  11. You’ll be able to’t use offscreen positioning so as to add labels (e.g., to types) that just some individuals, like customers of assistive know-how, can understand. All people has to see them.
  12. CSS layouts, notably these with absolutely-positioned parts which might be faraway from the doc circulation, might merely be prohibited on the highest stage. Actually, supply order should match presentation order even on the lowest stage.
  13. Additionally on the highest stage, it’s important to present a solution to discover all the following:
    1. Definitions of idioms and “jargon”
    2. Enlargement of acronyms
    3. Pronunciations of some phrases
  14. You even have to supply an alternate doc if a reader with a “decrease secondary training stage” couldn’t perceive your most important doc. (Actually, WCAG 2 repeatedly proposes sustaining separate accessible and inaccessible pages. In some circumstances, you don’t essentially have to enhance your inaccessible pages so long as you produce one other web page.)

Since these three paperwork are “drafts,” in fact all of the above can change. However actually, it received’t. A Final Name Working Draft is considered as considerably full. It’s “a sign that… the Working Group believes that it has happy its related technical necessities has happy important dependencies with different teams.” The WCAG Working Group isn’t going to budge on main points at this level.

It’s the definitions that sink it#section7

Whereas WCAG 2 requires all method of unrealistic and unproven options, these are usually not what’s going to sink the rules. One thing as mundane as definitions will care for that.

WCAG 1 was strongly HTML-specific. All people acknowledged that as an issue in an age when codecs that blind individuals like to hate, like PDF and Flash, are slowly turning into accessible. So WCAG 2 needed to be technology-neutral.

However in so doing, it imagined a parallel universe by which the overwhelming majority of net content material ceased to be plain-Jane HTML, CSS, and JavaScript. It envisioned a world by which heaps and plenty of Flash, PDF, and different, as-yet-uninvented codecs have been accessible and meant to be accessible. To accommodate this dreamworld, WCAG 2 was written and rewritten and rerewritten to use to every thing. Alongside the way in which, it misplaced the power to use to the actual issues actual builders work on daily—plain-Jane HTML, CSS, and JavaScript.

Pop quiz: What do the next phrases, given with their official WCAG 2 definitions, actually imply?

authored unit
set of fabric created as a single physique by an writer
authored part
an authored unit meant for use as a part of one other authored unit
net unit
a group of data, consisting of a number of sources, meant to be rendered collectively, and recognized by a single Uniform Useful resource Identifier (resembling URLs)
parsed unambiguously
parsed into just one information construction
programmatically decided
decided by software program from information offered in a user-agent-supported method such that the consumer brokers can extract and current this info to customers in several modalities

Are you able to translate any of those phrases into phrases that each reader of this text understands, like “web page,” “web site,” “legitimate,” “well-formed,” or “template”? Nicely, I can’t. Amid all these definitions, the place are the templates we use to create websites composed of legitimate, well-formed pages?

In the event you’re a standardista engaged on accessible web sites in the present day, are you truly, with out even figuring out it, an writer authoring authored items for use in authored elements in programmatically-determined net items that may be parsed unambiguously?

Check out WCAG 2 and also you’ll give you your personal guidelines of malapropisms and incomprehensible passages. Actually, a lot of WCAG 2 is so exhausting to know, and nearly unattainable to use to real-world web sites, that WCAG 2 is not any higher than its predecessor in a single respect—each paperwork flunk their very own tips for clear and easy writing.

If you can’t perceive the fundamentals of a suggestion, and if WCAG 2 usually is so aloof from the actual net that it may’t even trouble to make use of phrases that working builders perceive, are you realistically going to have the ability to implement WCAG 2 in your web site? Keep in mind, you can not formally fall again on the Strategies and Understanding paperwork for added info. Solely the WCAG 2 doc itself is “normative.” You sink or swim based mostly solely on that.

And if in case you have hassle understanding WCAG, does this not indicate that somebody might come together with a distinct interpretation and accuse you of violating WCAG, and, by implication, producing an inaccessible web site? Since that’s unlawful in some elements of the world, a sure diploma of readability is important, however readability is one thing you do not get in WCAG 2.

In the event you slog by means of WCAG 2, you’ll discover that even one thing as deceptively easy as that WCAG 1 guideline on clear and easy writing isn’t there. Neither is there something truly stronger than that guideline. Actually, there’s nothing in any respect alongside these strains to be present in WCAG 2’s Precept 3, “Content material and controls should be comprehensible.”

You do, nonetheless, should take fanatical care to mark up foreign-language passages, idioms, and the like, and in case your content material “requires studying skill extra superior than the decrease secondary training stage,” it’s important to present “supplementary content material” that doesn’t require that studying stage. In the event you’re a learning-disabled individual, that’s just about all WCAG 2 is keen to do for you.

Based mostly on my evaluation and on shows by Gian Sampson-Wild, evidently dyslexics and others with cognitive disabilities have been sacrificed on the altar of testing. As WCAG 2 tells us:

All WCAG 2.0 success standards are testable. Whereas some will be examined by laptop packages, others should be examined by certified human testers. Typically, a mix of laptop packages and certified human testers could also be used. When individuals who perceive WCAG 2.0 check the identical content material utilizing the identical success standards, the identical outcomes needs to be obtained with excessive inter-rater reliability.

“Excessive inter-rater reliability” isn’t outlined. Does it imply eight out of ten individuals? Six? All ten?

Plainly all people assumed it could be simple to seek out “individuals who perceive WCAG 2.0” but who additionally disagree {that a} sure phase of content material is clearly and easily written. I assume it was taken as axiomatic that checks of content material would seldom obtain “excessive inter-rater reliability,” which depends on messy human opinion. The Working Group was and is unreasonably fixated on automated testing, partly because of the presence on the Working Group of authors of automated testing functions and algorithms. The group was capable of abdomen the truth that, for instance, alt texts will be evaluated solely by people, however was unwilling to just accept that the identical applies to “content material” typically.

It’s harsh however truthful to look at that WCAG 2 sells out individuals with studying disabilities so {that a} instrument like Bobby, or a competing or successor instrument, can check a bigger variety of standards with the next success fee.

The inventive fiction of a number of ranges#section9

WCAG 1 had three ranges of “conformance,” which, in typical WAI model, got a complete of six names—Precedence 1/Stage A, Precedence 2/Stage AA (annoyingly written as “Double-A” to get round defective screen-reader pronunciation), and Precedence 3/Stage AAA (“Triple-A”). Standardistas ultimately found out that Priorities 1 and a couple of have been what you actually wanted to make an accessible web site; Precedence 3 was strictly non-obligatory (additionally onerous and unattainable to fulfill in precept). Even some governments, like Canada’s, require Precedence 2 compliance for their very own websites, although it’s not essentially achieved.

When specialists perform evaluations of internet sites towards WCAG 1, more often than not they contemplate the primary two precedence ranges. Few, if any, websites go Precedence 3 analysis; the Incapacity Rights Fee and Nomensa discovered that no websites examined met Precedence 3.

To a rational observer, all which means that Priorities 1 and a couple of in WCAG 1 are actually a single algorithm and Precedence 3 is irrelevant and unattainable. Getting this concept by means of the heads of the Working Group (or slightly, by means of the pinnacle of one of many cochairs) was unattainable, so in WCAG 2 we’re nonetheless caught with three ranges. However get this: All ranges are deemed vital.

    Stage 1 success standards:

  1. Obtain a minimal stage of accessibility.
  2. Can fairly be utilized to all net content material.
    Stage 2 success standards:

  1. Obtain an enhanced stage of accessibility.
  2. Can fairly be utilized to all net content material.
    Stage 3 success standards:

  1. Obtain extra accessibility enhancements.
  2. Can’t essentially be utilized to all net content material.

To translate: We poor saps misunderstood WCAG 1’s precedence ranges to be actual precedence ranges. WCAG 2 considers all of its tips “important for some individuals,” although they’re nonetheless damaged up into three ranges. However truly, in the event you look carefully on the WAI paperwork:

  1. Even in the event you adjust to all three ranges in WCAG 2, you should still find yourself with an inaccessible web site.
  2. You by no means should adjust to greater than half of the Stage 3 tips.
  3. The WCAG 2 doc itself baldly states that “It isn’t really useful that Triple-A conformance ever be required for whole websites.”
  4. In a round contradiction, Guideline 4.2.4, at Stage 3, doesn’t even require you to fulfill Stage 3 in some circumstances.

Which stage would you want to adapt to? Please make your choice now.

In an additional absurdity, the Working Group couldn’t even finesse its tips to use to all ranges. Some tips don’t even manifest themselves at Stage 1, the bottom stage. I did a rely:

  • Ranges 1 + 2 + 3: 7 tips
  • No Stage 1: 1 guideline
  • No Stage 2: 2 tips
  • No Stage 3: 1 guideline
  • Stage 1 solely: 2 tips
  • (Stage 2 solely or Stage 3 solely: Nil)

It’s as if net requirements by no means existed#section10

Whereas individuals such as you and me have been labouring within the trenches since roughly 1998 to enhance net requirements—enhance help in browsers, enhance understanding amongst authors, enhance the essential process of explaining requirements—the WCAG Working Group has been off in its parallel universe cooking up tips that apply equally ambiguously to every thing. However the Working Group definitely did take the time to exterminate some accepted ideas.

Sure, we all know already: A web site with legitimate HTML isn’t mechanically accessible. We’ve received a few enjoyable little instance pages to have a look at (by Gez Lemon and Bruce Lawson). However that’s all they’re—examples. In the actual world of clueless tag-soup builders, the rising minority who perceive legitimate HTML are an elite who additionally perceive accessibility. They perceive which accessibility options you get without cost with legitimate HTML (like alt texts, which—sure, we all know already—should be written accurately). These builders take the time to incorporate the remaining accessibility options anyway.

In addition they perceive that tag soup produces unpredictable ends in browsers and in display readers. They know {that a} single unencoded ampersand, or omitted semicolon, or stray Unicode character on a web page might knock it into the land of invalid HTML, however these are trifling examples not present in tag-soup websites like Amazon and eBay. (They know that Amazon and eBay are profitable regardless of their supply code.) They know that validity is a fragile factor that certainly can be blown out of the water by one thing so simple as a personality like an é, an &nbsp (sic), or an & within the improper place. They know all that.

Nonetheless, legitimate HTML was a second-level requirement in WCAG 1. You nearly by no means discover it in a industrial web site—Nomensa’s current survey, which discovered 4 examples out of 99 websites it manually checked, is the best I’ve ever seen. However, as a requirement, it warned builders that, whereas tag soup is the norm, it’s not what we need.

WCAG 2 upends that apple cart fully. You by no means should have legitimate HTML in WCAG 2–compliant websites. All that’s required is that the web page be parsed unambiguously (Guideline 4.1—a Stage 1 guideline with no Stage 2 or 3 tips). That is purported to imply “no improperly-nested parts,” however you’d by no means know that from the time period itself.

In an HTML web page, you’ll be able to hold proper on utilizing all of the misplaced stray characters you need, however you’ll be able to’t nest <p> inside <p>. You would not have to make use of any parts or attributes required by the specification. You would not have to make use of parts in keeping with specification. All this spells hassle for the case of types, an space of fixed annoyance for screen-reader customers. A doc made up of nothing however divs and spans, if unambiguously parsable, passes WCAG 2 free and clear.

XHTML pages, in keeping with spec, are purported to cease useless of their tracks on the first ill-formed content material, however we all know they don’t accomplish that in the actual world, the place XHTML is handled like a sort of HTML with added closing slashes
(save for the tiny few perfectionists who serve XHTML as XML). So in reality this requirement offers XHTML the identical go it offers HTML.

Does any of that basically remedy the issue? Or does it have sufficient of an look of fixing the issue that it may very well be voted into existence by Working Group members from firms like IBM, Oracle, and SAP, whose software program can not reliably produce real legitimate HTML? (IBM has been actively selling a DHTML accessibility method that breaks the HTML spec. Oddly, and futilely, the Strategies doc discourages such a factor.)

Do you suppose WCAG 2’s guideline is sweet sufficient to enhance the practices of tag-soup builders? Even when legitimate HTML all over the place on a regular basis is unattainable, the actual fact stays that, in 2006, we have now by no means had extra builders who perceive the idea and are attempting to make it actual on their very own websites. WCAG 2 undoes a requirement that, have been it retained, may very well be completely timed now.

Captioning and audio description for multimedia#section11

If there’s any space by which the software of WCAG 1 was a complete failure, it’s multimedia. Folks have been fairly pleased to disregard the necessities for captions (for the deaf) and audio descriptions (extra narration for the blind), each of which have been required on the lowest accessibility stage. (Truly, it was worse than that from a deaf individual’s perspective. You may get by simply with a transcript, not precise captions.)

Captioning and outline merely are usually not discovered within the wild. When there’s any entry in any respect, it’s by means of captioning. On this method, on-line multimedia follows TV, residence video, and cinema in main democracies, the place captioning is frequent and outline isn’t. (Who can overlook the irony of AOL’s head of accessibility, a blind man, asserting captioning on “choose” AOL movies, however no audio description in any respect?)

For a deaf or a blind one who desires to know multimedia, WCAG 2 presents no actual enchancment. The transcript-only loophole has been closed, and captions stay a requirement on the lowest stage for prerecorded video. However as an alternative of audio description, you may get by with a figment of the Working Group’s creativeness referred to as a “full multimedia textual content various together with any interplay”. A discredited holdover from WCAG 1, it’s apparently a mix of transcript of dialogue and sound results (which blind individuals don’t want), transcript of audio descriptions (which deaf individuals don’t want), and hyperlinks to any interactive elements within the video.

The entire thing is meant to be of assist to deaf-blind individuals, who have been by no means surveyed for his or her preferences, an motion I really useful to WAI at a face-to-face assembly in 2003. Nor was any consumer testing carried out. (That’s all I can inform from revealed proof, anyway. I despatched e-mail inquiries to deaf-blind organizations in a number of international locations asking in the event that they’d been surveyed or had any opinions, with no response.)

There are about three recognized examples of such a transcript within the seven-year historical past of WCAG (e.g., DigNubia). And there actually aren’t any HTML semantics for such transcripts, except you wished to push the envelope of the definition record (a banned utilization in “HTML 5”).

On the next-to-lowest compliance stage, all of the sudden actual audio descriptions are required and, once more all of the sudden, dwell video should be captioned. Go one step increased and it’s important to translate your video into signal language (which one?) and supply that very same imaginary transcript, amongst different issues. You by no means have to explain dwell video.

And whereas I’ve by no means been a proponent of requiring the lots of of dwell on-line radio stations to caption themselves, definitely prerecorded podcasts are an apparent supply of inaccessible multimedia. However truly, multimedia is outlined as “audio or video synchronized with one other sort of media and/or with time-based interactive elements.” Your MP3 podcast isn’t synchronized with something, so it’s exempt. This requirement will fulfill nearly all of podcasters who ever even bothered to consider accessibility, just about all of whom determined it was an excessive amount of hassle even when they preferred the thought or labored for WAI on the time. The requirement will even be sure that the established order of inaccessible podcasting stays untouched.

That’s sufficient for one article, I believe. However that isn’t the top of my feedback on WCAG 2; you’ll be able to test my web site for ongoing additions. This text’s feedback part, and the tag WCAG2, are different methods to remark.

Saying the WCAG Samurai#section13

WCAG 2 isn’t too damaged to repair, however we have now no motive to suppose the WCAG Working Group will truly repair it. The Working Group is simply too compromised by company pursuits, too wedded to the conclusions we see within the present “draft,” too damaged usually. What you see in WCAG 2 now’s just about what you’re gonna get—completely.

As such, WCAG 2 will probably be unusable by real-world builders, particularly standards-compliant builders. It’s too obscure and counterfactual to be a dependable foundation for presidency regulation. It leaves too many loopholes for builders on the hunt for them. WCAG 2 is a failure, and never even a noble one at that.

If that is what we get when WAI tries to rewrite WCAG from scratch, perhaps there’s an alternative choice. WCAG 2 doesn’t “exchange” WCAG 1 any greater than XHTML “changed” HTML. Possibly all we actually have to do is to repair the errata in WCAG 1. It’s been mentioned earlier than, however a WCAG 1.0 Second Version or a WCAG 1.1 by no means occurred.

Now, although, I can announce that such errata actually are going to be revealed, and my buddies and I are going to do the publishing. After the way of Zeldman’s CSS Samurai posse, which put CSS layouts on the map for browser makers and builders, the WCAG Samurai will publish errata for, and extensions to, present accessibility specs.

In fact we aren’t going to infringe anyone’s copyright, however one other factor we’re not going to do is run a completely open course of. It’s a viable mannequin for requirements growth, one I’ve championed in one other context, however in net accessibility it’s confirmed to not work. Membership in WCAG Samurai, as in CSS Samurai, will probably be by invitation solely. If we would like you, you’ll hear from us.

In fact that is unfair to say the least, if not actively elitist and hypocritical. Name it as you see it. However that is what we’re going to strive within the hopes of getting the job accomplished, which WAI and its mannequin have merely didn’t do.

Leave a Comment