Comic Update: HTML5 Stubborness and Snogging
Posted by Kyle Weems on June 15, 2009Today’s comic references two very important topics that everyone should know about.
The first involves Bruce Lawson and snogging. In relation to point #2, I tweeted this. He responded with this. I find the word snogging hilarious, so it went downhill from there, with mental images of Ian Hickson and John Foliot getting hot and heavy.
In those mental images, Ian is asked to shave.
The second topic, which quite inadvertently spawned the first, involves HTML5, ARIA and the apparent lack of peace between the groups responsible for developing each. In his post Alternate Text in HTML5, Bruce bravely discusses his opinion on the topic despite his stated delicate nature and then suggests a group hug, and perhaps a sing along.
By comment #2, Anne van Kesteren has dropped the thunder and brought back the fighting.
Here’s a recap: Blind people can’t see. Blind web users, as a result, need some aids to make sense of things we’d take for granted, even when screen readers are taken into account. Pictures need some form of alternate text and tables need some sort of summary to help give them the scope of the data that’s about to be read to them (as just two examples.)
The WAI-CG has methods for solving these sorts of problems. These solutions exist in HTML4. However, the WHAT WG, with what I presume is a desire to keep code simple, want to do accessibility their own way. To prove their point, they lean on surveys of existing web content which show little adoption of the accessibility features being debated. They also decline to accept the advice of accessibility experts with real-life experience interacting with disabled users.
For a bunch of smart people, that’s pretty stupid.
Actually, that’s stupid for stupid people, so it’s outright dead-brained for smart people.
Why would surveys of existing content prove the effectiveness of the features when used? All it proves is that accessibility awareness needs to be raised among developers. To figure out whether the proper use of these features improve accessibility for the blind, I’d suggest talking to a blind web user.
As John Foliot points out in his comments in Bruce’s post, by all accounts Ian has not actually received any input from a blind person on the accessibility features he is denying.
I’m not an expert in this field, so I’m not going to propose solutions. I do propose, however, that the WHAT WG listens to the experts instead of continuing to cling to their “not invented here” mentality and looking to their own interests before those of the community that absolutely relies on accessibility to make use of the web.
In other words, stop being jerks.
Here’s a couple of related links to the topic in addition to those shown above that might make a good read: Mechanism to Summarize a Table, maintained by Laura Carlson. HTML5 and WAI-ARIA by Anne van Kesteren (the real good stuff is in the comments). Also, make sure to check out the comments in Bruce’s post. There’s a lot of good material in there to get a feel for positions and justifications.
Edit: Corrected the authorship of the Mechanism to Summarize a Table link, based on John’s correction below. Sorry for that, Laura!
Tags: accessibility, alt, anne van kesteren, aria, bruce lawson, html5, john foliot, not invented here, snogging, summary, wai cg, what wg
Kyle – credit where credit is due. The wiki page: “Mechanism to Summarize a Table” was not actually written by me (although there is some content in there that I contributed), but rather is painstakingly kept current by Laura Carlson. I (like others) believe that it is a fair and balanced review of the issue, the current solutions (warts and all), & some new ideas (good and bad). Unlike much of the behind the scenes discussion that the WHAT WG partake in, this is a publicly available document that any and all can access at any time. It represents what real open dialog should look like (IMHO). (The wiki also contains entries such as “Omitting Text Alternatives on “, and “Table headers attribute Issue”)
As for snogging, we’ll leave that kind of behavior to Mr. Lawson’s creative mind, however I can point to: http://tinyurl.com/n36c33 to perhaps calm your mental images.
And Bruce… JF prefers Stevie Ray Vaughn and John Hammond Jr. over the New Seekers in a heart-beat.
“They also decline to accept the advice of accessibility experts with real-life experience interacting with disabled users.”
No. Let’s be clear on this: they *disagree* with *some* accessibility experts. Opinions vary on the best way to provide accessibility features. See for example: http://blog.whatwg.org/the-longdesc-lottery
Your article is misleading, because it implies the WHATWG are rejecting expert opinion. They’re not. They’re advocating *other* expert opinion. This topic is fraught enough (for reasons I don’t fully understand) without any further misrepresentation or name calling.
Laura Carlson has done an amazing job of capturing all of the information on these topics and organizing it into one detailed page.
mattur, who are the *other* accessibility experts, though? All of the accessibility experts are more or less on the same page, from what I can see. Some differences of opinion on some implementation issues, but more or less agreed on what should, and shouldd not, be in HTML5.
Surely you don’t think Ian Hickson qualifies as an accessibility expert, do you?
@Shelley & John – Thanks John for the correct authorship information. The more I hear about Laura Carlson, the more I like.
@mattur – Your linked post describes a sampling of content indexed by Google, as reported by Ian. Correct me if I’m wrong, but neither a cross-section of web pages nor Ian Hickson qualify as accessibility experts. You can’t simply look at the “expert” part of the title and assume that qualifies him to be a reliable source on any topic.
No matter how good I am at doodling squirrels I shouldn’t be performing heart surgery.
@Kyle: The author of that post is Mark Pilgrim of DiveIntoAccessibility fame, who previously worked as an accessibility architect at IBM. Check the comments:
The RNIB and WebAIM both recommend not using longdesc, making the content available to *everyone* instead:
http://blog.whatwg.org/the-longdesc-lottery#comment-9044
Joe Clark wrote the well known Building Accessible Websites book:
http://blog.whatwg.org/the-longdesc-lottery#comment-9051
Tomas Caspars shares his experience running a large site focused on people with disabilities:
http://blog.whatwg.org/the-longdesc-lottery#comment-8947
http://blog.whatwg.org/the-longdesc-lottery#comment-9127
And incidentally, legendary sex-god Bruce Lawson also prefers accessibility features that are available to *all*, not just AT users. And FWIW I’ve been building accessible websites since the mid-90s, pre-WCAG1, and I agree with him.
None of this shows longdesc or other AT-only features are wrong or sub-optimal. It just shows expert opinions differ. There are some almighty egos clashing here; don’t let the level of noise from some persuade you that one opinion is “right” and the other “stupid”. Because that would be, er, stupid. And wrong.
I must respectfully disagree with Mattur here. His solution, Kyle, to the question you posed to me on twitter would be that all of the dialog contained in your cartoon should be included as “in the clear” text on the same page of as the cartoon. Go, read his comments and Mr. Pilgrim’s… that is their proposed solution.
The fact that this flies in the face of reality seems to be missing: will you, as an artist that is creating graphic images that contain long tracts of text also include this in the clear text? Hardly. So what then?
I find it interesting as well that the only source of dissent Mattur can offer is posting to one blog entry that is now almost 2 years old (yes, this same question has been going ’round and ’round for over 2 years). For what it’s worth mattur, I’ve come to accept that there are other ways of linking the text content to images outside of longdesc, but still consider it foolhardy and thick-headed to eliminate the accessibility feature from HTML5 based simply on the say so of Mssrs. Pilgrim and Hickson.
The real question however is on ‘credability’ of expert opinion:
Yes, once upon a time Mark Pilgrim* worked in the accessibility field at IBM… surely then he could get former associates and distinguished accessibility experts such as Richard Schwerdtfeger [http://tinyurl.com/24ffw3] to comment and confirm their *opinions* regarding the best way forward. Notably absent. Perhaps TV Raman [http://tinyurl.com/njened or http://tinyurl.com/myxlak who also once worked at IBM, Adobe and is now part of Google’s web accessibility team (ya know, where both Hickson and Pilgrim now work), and who, BTW, is also a blind user… surely *he* could dispute or refute the claims that accessibility experts have been consulted – they have not. What about Jim Thatcher, who essentially authored the Section 508 guidelines, and was one of the expert witnesses called in the Target.com accessibility case? (nope) NFB, RNIB, Section 508 board, *any* official entity that supplies services to disability groups (not just blind either, as accessibility transcend that one type of impairment)… no, despite repeatedly asking for a name of *any* professional paid to deliver accessibility and technology to end users, none has been submitted. So again Mattur, can we have some names please?
@John & mattur – Just so I understand, the solution of text “in the clear” that Pilgrim and Mattur propose in his linked post is placing all the descriptive text of the image on the page itself? That seems clunky, design unfriendly and in many cases downright distracting for sighted web users.
@mattur, it is disingenuous to carry on about longdesc as if the W3C WAI CG consensus decisions (http://www.w3.org/2009/06/Text-Alternatives-in-HTML5) about short and long text alternatives continue to advocate its use. It is in fact the opposite, it says
“we believe it is acceptable to obsolete longdesc in HTML5.”
It instead advocates the use of aria-describedby to provide a programmatic association between the markup of a longer text alternative (if needed) whether it be provided on the same page or a different page. No content is hidden and it does not say that the programmatic asscociation needs to be hidden. all it is saying is please identify that content x is a long description of image y, because implicit visual association is not enough.
What is the issue with this? the use of aria-describedby does not involve in any information loss, it provides extra information that any user agents can convey to the user.
@John Foliot: anyone who disagrees with you, like for example the RNIB or WebAim or Joe Clark, has no credibility. Therefore, there’s only one valid, credible opinion, which just happens to be yours. Got it.
@Kyle: Other options are available e.g. use a standard link on an image to provide a longdesc content page, come up with a better website design, realise the web is more than just pretty pictures etc etc.
@Steve: Yes, I was just using longdesc as an example of differing opinions. I think using aria-describedby where applicable is a better solution.
Ironically, the WAI CG basically agreed with those meddling kids, erm I mean the HTML5 cabal, on obsoleting longdesc and making alt optional.
I look forward to JF posting vicious comments about the WAI CG’s lack of credibility, and Kyle calling them all “stupid”.
Not being vicious, simply stating facts:
1) Methods of Providing a Long Description
There are several ways of providing a long description for images. These options are listed below, in order of most preferred to least preferred:
1. Provide the long description in the context of the document itself
2. Provide a link to a long description via a normal text link
3. Provide a link to a long description via the longdesc attribute
4. Provide a link to a long description via a “d” link
[Source: WebAIM - http://tinyurl.com/qp4hv - Mattur, please note #3]
***********
2) Complex images are images whose full meaning cannot be adequately described in a short phrase or sentence. This may include graphs, charts and maps.
A brief name or description should be given in the ALT text, and a longer description of the content of the image given elsewhere. There are a variety of ways in which this can be achieved. A text description can be given on the same page, or linked to, from the image or the page. The ALT text for the image could then be “Office layout, see description below”. LONGDESC can also be used but note that not all browsers use the “longdesc” attribute. Finally, a “D” link can be used.
[Source: RNIB - http://tinyurl.com/l6o5zw - again Mattur, please note suggestion #3]
***********
“The true accessibility method for a complex image may be a tactile graphic.” (Say what? – JF)
[Source: Joe Clark - http://tinyurl.com/25a382 ]
“Quite simply, I want it all, and so should you. Give us everything you’ve got. Give us everything there is to give.”
[Source: Joe Clark - http://tinyurl.com/lvk38n ]
“Waxing poetic: Long descriptions
When alt and title are insufficient to embody the visual qualities of an image in words, HTML offers an access feature at a higher level of complexity: longdesc.
A longdesc is a long description of an image. How long? The sky’s the limit. longdesc is meant for pictures worth more than a thousand words (or 1,024 characters). The aim is to use any length of description necessary to impart the details of the graphic. It would not be remiss to hope that a long description conjures an image – the image – in the mind’s eye, an analogy that holds true even for the totally blind.”
[Source: Joe Clark - http://tinyurl.com/yhv9ex ]
All valid, credible opinions to me Mattur.
And besides, you miss the point: You claim that “…they *disagree* with *some* accessibility experts. Opinions vary on the best way to provide accessibility features….” And so, once again, which accessibility experts *DO* agree with the HTML5 WG on their vision of the “best way to provide accessibility features”?
CAN WE HAVE SOME NAMES PLEASE?
Regarding the WAI CG – I fully endorse the latest recommendation put forth by the WAI CG to the HTML5 WG on their “Consensus Resolutions on Text alternatives in HTML 5″ [ http://www.w3.org/2009/06/Text-Alternatives-in-HTML5 ] – it remains to be seen whether the HTML5 WG will accept this expert guidance, given that it *is* the culmination of the hard work of 4 different W3C Accessibility Working Groups – groups that include technologists, advocates, and DISABLED USERS. Or will it be received with typical WHAT WG arrogance: http://krijnhoetmer.nl/irc-logs/html-wg/20090611 (posted without further comment)
@mattur – I believe I called the ideas stupid and referred to the people as smart. I’m aware that’s splitting hairs, but that’s more consideration than many WHAT WG members have shown to others in IRC. I don’t plan on backing down on my comment that using markup statistics to back a proposition on accessibility markup’s effectiveness isn’t a smart idea. Everyone in this conversation is fully aware that a large portion of the web isn’t being authored by standardistas or accessibility-aware developers, so they shouldn’t be relied on as experts of good methodology.
John Foliot: If I had spent the past two years insulting and disparaging anyone involved in HTML5 who suggested making longdesc obsolete and alt optional, and then “…the hard work of 4 different W3C Accessibility Working Groups – groups that include technologists, advocates, and DISABLED USERS” recommended exactly that, I’d probably be feeling a bit foolish too.
Mattur,
You once again make claims based upon your perception… please back up your claims with URLs will you? I hardly feel foolish – it is worth pointing out that while hangers-on like you deem to take pot-shots from the sidelines, the current proposed solution for alternative text for images [ http://www.w3.org/2009/06/Text-Alternatives-in-HTML5 ] is in fact based upon an idea that I first articulated in an email, as documented here: http://esw.w3.org/topic/HTML/IssueAltAttribute (Now who looks foolish?)
What I *have* spent the last two years doing is shining the light on the fact that the HTML 5 Working Group has made sweeping claims and changes to items specifically related to web accessibility, without properly consulting the affected community and those who work directly in that community. They have wholesale removed existing attributes from HTML 4 that were created to enhance accessibility without providing suitable replacements, all the while continuing to support presentational artifacts such as the bold and italic ‘tags’, and introducing half-baked new elements such as canvas that have zero accessibility considerations included. (an identified and acknowledged ‘problem’) Despite repeated requests to share with the web accessibility community both the hard data that they allegedly base all of these decisions on, as well as multiple requests to share with us *whom* they have in fact consulted with, neither has been provided – what we see instead is a small group of non-disabled technologists ‘deciding’ the best solutions to existing problems and dictating to the affected communities “this is how it is” (or in the case of canvas, simply not thinking about or addressing the potential problems that were obvious to the web accessibility community the first time they saw canvas). It is worth noting as well, that it is not only the web accessibility community that has issues with the HTML5 processes and procedures, but also other communities such as the RDFa community, and to a lesser extent SVG community.
I make no apologies for being a vocal and outspoken critic of this process, and of many of the ill-informed decisions that this small group makes: however, in every instance I back my claims with 3rd party URLs, studies and external documentation – I *never* simply insist that others take *my* word for it. Mattur, can you say the same? While I quote facts, you are reduced to calling me names. Got it. (And, I will presume, so do the other readers)
John Foliot:
“…[people] might be surprised to hear of some of the sillier regressions that HTML 5 is suggesting, such as making the ALT attribute optional, the deprecation of accessibility enhancements such as @longdesc…”
http://meiert.com/en/blog/20080708/you-can-use-html-5/#comment-70942
The WAI CG recommended making alt optional and obsoleting longdesc, as in the current HTML5 draft.
So either expert opinions vary on the best way to provide accessibility features, or you’re not an expert, or the WAI CG aren’t experts. Which is it?
Yes, but I would caution against digging up old-timers’ opinions and presenting them as those old-timers’ current opinions. In the clear light of day, it is apparent to me that LONGDESC was barely ever used because images complex enough to require it are barely ever published. So if you’d all like to act as though I have some kind of horse in this race, which I don’t, at least saddle up the right horse.
The idea of refusing to require an alternate text for an image is nonsense, of course. We now have several methods beyond ALT, and even in the era of HTML 4 we had more options even than just ALT.
Why do I bother with you Mattur? What practical contribution have you made to HTML5 beyond sideline cheerleader and hixie fanboy?
Yes, a year ago I wrote that deprecating longdesc was silly – I still think that way (as I believe the attribute is not flawed, but rather we’ve done an extremely poor job teaching proper implementation), but since that time the overall web accessibility community has discussed practical alternatives to longdesc (aria-describedby); conversations that I have been a contributor to BTW, and subsequently a new “community” position has emerged based upon those discussions – the one now articulated in the W3C CG document, which I’ve already publicly endorsed (a year after that ’smoking gun’ blog posting of mine that Mattur pointed to).
However, it should also be stated that the WAI document attaches some *specific conditions* to the deprecation of longdesc:
* IF
o aria-describedby is incorporated in HTML5
o and aria-describedby allows pointing to long text alternatives that are off of the page (by pointing to a link on the page)
* THEN
o we believe it is acceptable to obsolete longdesc in HTML5.
[ http://www.w3.org/2009/06/Text-Alternatives-in-HTML5 ]
This is not my personal best option, but one that is palatable to me, so long as the IF conditions are met *as described*. It is also represents a compromise position (at least for me personally) based upon informed discussion and debate amongst web accessibility experts and professionals whom I respect, as well as end users directly affected by this recommendation. It is also in marked contrast to how the WHAT WG cabal have rolled out their non-debatable ideas on ensuring “accessibility”…
(Meanwhile, at least one WHAT WG cabal member states: “I don’t see aria-describedby becoming much of a success – It seems way too complex for everyday use” [ http://krijnhoetmer.nl/irc-logs/html-wg/20090611#l-69 ]: apparently <img src=”" alt=”short alternative” aria-describedby=”path to longer description” /> is too ‘complicated’ for some WHAT WG members… )
SO, for absolute clarification, as Mattur seems to have a real hard grasping these things: I believe that longdesc should not be deprecated, but instead should continue to exist, while a newer (perhaps better) method be introduced (aria-describedby). This idea BTW, is one that I share with my friend (and Opera’s Chief Standards Officer) Charles McCathieNevile:
“It seems to me that we should be far less hasty to remove things from HTML which were designed to support accessibility – and far more hasty to work on adding an alternative which is better, in order to actually field test the two side by side for long enough to understand which of our assumptions were brilliant insights and which were foolishness…”
[ http://lists.w3.org/Archives/Public/public-html/2009Jun/0275.html ]
HOWEVER, despite my personal reservations, I continue to support and endorse the W3C CG document when it comes to “Recommendations regarding Long Text Alternatives” as I recognize it as a consensus document arrived at by my peers, and it is a consensus position that I can live with.
I’ve already discussed my position (and contributions) to @alt earlier, so what point then are are you ultimately trying to prove? That I (unlike others on the other side of the discussion) am actually less dogmatic than the WHAT WG Cabal, who insist that they have a lock on the one truth, and all others are infidels? ‘Cause that sure seems to be the case here: I’ve moved, they won’t. That John Foliot can accept the collective wisdom and debate/discussion of other Web Accessibility experts and actually move on his initial opinion? (unlike the WHAT WG – who refuse to accept *any* expert opinion from outside their small group of buddies?) Guilty as charged! (like this is a bad thing?)
Anything else you are unclear on Mattur?
John Foliot:
“I’ve long believed that when your adversaries cannot attack your position, they attack your character…”
http://john.foliot.ca/sticks-and-stones-can-break-my-bones/
Mattur is right, I called him names: sideline cheerleader and hixie fanboy
That was wrong and I apologize.
I look forward to hearing what Mattur’s actual positions are, beyond appearing to whole-heartedly agree with the WHAT WG when it comes to ignoring expert opinion, rejecting consensus, and thinking that “…aria-describedby is too complex…”
Mattur:
* Do you support and endorse the WAI CG document?
* Do you believe that the WAI CG document should be the final word in the issues of long and short text alternatives in the Draft HTML 5 document?
* Do you support and endorse the W3C process of consensus when it comes to creating the next generation HTML?
* Do you endorse or reject Ian Hickson’s statement: “The HTML5 work isn’t using the traditional W3C approach, and will never use a consensus approach so long as I am editor.” [ http://lists.w3.org/Archives/Public/public-html/2008Jul/0354.html ]?
* When appropriate, will *you* use aria-describedby to label your complex images, or do you too find it “too complex” for daily use?
Mattur, for clarity my answers are yes, yes, yes, reject, and yes.
@John Foliot said on June 16th, 2009 at 5:23 pm:
>>>>I must respectfully disagree with Mattur here. His solution, Kyle, to the question you posed to me on twitter would be that all of the dialog contained in your cartoon should be included as “in the clear” text on the same page of as the cartoon. Go, read his comments and Mr. Pilgrim’s… that is their proposed solution.>>>>
@Kyle Weems said on June 16th, 2009 at 6:12 pm:
>>>>@John & mattur – Just so I understand, the solution of text “in the clear” that Pilgrim and Mattur propose in his linked post is placing all the descriptive text of the image on the page itself? That seems clunky, design unfriendly and in many cases downright distracting for sighted web users.>>>>
Kyle, there is no problem with including long descriptive text inline; css can be used to hide it so that sighted readers don’t trip over it, whilst non-visual readers, who don’t currently depend on css, will hear it via their screen reader or read it via their Braille display.
This one, sadly, seems to be one where the heat created by the controversy is greater than the factual content of the controversy.