Rebuilding The Web

1. Posted by mattur
on Monday 2011-01-17 at 12:23:09 PST

Yes

2. Posted by Lucica
on Monday 2011-01-17 at 14:49:50 PST

It's a sad news. I think W3C did not learn anything from the past 10 years' experience. It does not matter if it's CSS, HTML, A11Y related or anything else, reading their documents it's like reading the Constitution. Which means their own A11Y information for example is not accessible to all the people due to over-complicating the documentation, the sitemap and the language. I do not intent to criticize W3C, their work is much appreciated and we are all grateful for their effort, but what is the use of it if it does not reach the masses? Who is implementing all this? :(

3. Posted by steve faulkner
on Monday 2011-01-17 at 15:12:19 PST

@vlad
here's a version of the alt doc with all the pics and code examples and supplementary material, references taken out http://bit.ly/e1dPTM (5 pages)

vlad wrote
"will also become the basis for future derivative works such as articles, tutorials, and references."

That is the point of providing the examples and references and glossary (whucih make up 40 of the 45 pages) and trying to cover all use cases, if derivatives are going to be made from the source material not making it as comprehensive as possible would be wrong.

@lucia the W3C are not other beings, if you have feedback about the alt document feel free to comment or write a bug.

I would take your feedback more seriously vlad if you didn't over simplify the reality.

5. Posted by Iza Bartosiewicz
on Monday 2011-01-17 at 17:07:19 PST

I think Dey Alexander's decision tree is the most friendly way of explaining text alternatives that I know of (link below). In most cases, adding appropriate and meaningful text alternatives will be straight forward. However, I appreciate the variety of examples given in the W3C document discussed here, because they cover those not-so-obvious cases. Reading them all in one go would be time-consuming, but they will be invaluable as a reference.

Text alternatives for images: a decision tree
http://www.deyalexander.com.au/blog/2010/12/text-alternatives-for-images-a-decision-tree/

6. Posted by Denis Boudreau
on Monday 2011-01-17 at 20:12:06 PST

If "perfect" alt text were as easy to implement as you say they are Vlad, authors wouldn't be screwing it up so badly. Truth is, there are no simple answers and there is a need for such a document. Maybe it's not meant for everyone, but its relevant nonetheless.

8. Posted by steve faulkner
on Tuesday 2011-01-18 at 02:24:50 PST

@vlad
as it says in the alt doc (http://dev.w3.org/html5/alt-techniques/):

"All feedback is welcome.
If you wish to make comments regarding this document, please submit them using the W3C public bug database. If you cannot do this then submit them to public-html-comments@w3.org"

"This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress."

10. Posted by mattur
on Tuesday 2011-01-18 at 10:26:53 PST

> "Here are stats that show how skewed the bug submission processed is against accessiblity issues:"

Can you think of any other reasons why the proportions displayed in the charts may differ?

11. Posted by steven faulkner
on Tuesday 2011-01-18 at 10:40:28 PST

@vlad I am not a W3C member, the company I work for is not a member, I have been involved in the accessibility effort in the HTML WG since 2007, it has been and continues to be hard work, but I would rather be involved in trying to shape the web and open standards via the W3C consensus process rather against bleating from the sidelines like you.

13. Posted by steve faulkner
on Tuesday 2011-01-18 at 13:00:51 PST

vlad wrote:
"when you build consensus amount W3C members, do you insult them as well?"

I was merely responding to the snearing tone of your previous remarks:

vlad wrote:
"One version for people with limited time to waste and the other for those who need a sedative to fall asleep?"

14. Posted by Denis Boudreau
on Wednesday 2011-01-19 at 21:04:46 PST

Vlad wrote: "1. Some of the things the 45-page document recommends cannot be implemented by WYSIWYG/authoring tools - the rules are too complex for implementing in a UI."

That may be true, but how is this relevant? If some things are harder to implement, they're harder to implement. Period. Wishing them to be simpler or oversimplifying like you do here will not change how complicated these "rules" really are. You suggest the following as a "perfect rule" for alternative text:

Vlad wrote: "That's simple too: you read the text that replaces the image within immediately surrounding content, and if the text fits into the flow of surrounding content, you have perfect alternate text."

I'm sorry, but this is simply not true, incomplete, inaccurate and (with all due respect) probably shows how short-sighted you are when it comes to alternate text. Alt text is more than assigning a value to an attribute. Among other things, it's pondering whether an image needs alternative text in the first place and if it does, considering whether the value of the text alternative is supposed to represent the image itself, or it's purpose, which is a totally different ball game.

Let's consider the two following examples:

Let's say the submit button for the search engine on my website is an image of a white triangle (pointing right) in a red rectangle. Classic stuff. What should my alt text be for that image? Does your perfect solution address that case? Nope. Because it doesn't explain to authors that they can deviate from systematic description to represent purpose instead. The expected alt for most people would be to describe the image (white triangle pointing right, on a red background) while in this particular case, what the user needs is a description of it's purpose (search the website). Your perfect rule doesn't cover such a case.

Let's say I have a picture of president Obama doing a speech on a webpage included in an article presenting his latest speech. Most authors would fill in the alt text for the image with something like: picture or image of President Obama giving a speech. However, as you most probably know, it is pointless (not to mention repetitive) to begin an alternative text with "picture" or "image" as screen readers will announce it. Most authors would fall in that trap and again, your perfect rule doesn't quite cut it.

There could be a lot more examples, and this is one of the reasons why Steve's document takes up 45 pages at the moment. I would even expect it to grow over 45 pages, and I totally trust his judgment in doing so because I know he is looking at every little details (unlike you) while developing it.

Vlad wrote: "2. Do you think authors of derivate works are going to accurately abridge and condense the 45 pages into a 1 or 2-page article or tutorial?"

Of course they will. We (the community) are already doing so with much more complex standards- or web-related subjects. Some will do it properly and others will try and do a poor job at it.

And on the other hand, maybe the subject of alternative text is complex enough to justify more than one or two 1 or 2 page articles or tutorials in the first place. I would expect sites like alistapart, thinkvitamin or sitepoint to get it right. Among others.

Vlad wrote: "3. Will this document make the browser vendors correctly implement alternate text rendering?"

If the browser vendors really wanted to take a stab at this and implement it correctly, they would yes. They have the brains and resources to do so.

15. Posted by Denis Boudreau
on Wednesday 2011-01-19 at 21:05:25 PST

You end your article with this plea:

Vlad wrote: "If you truly want to make the Web accessible, then focus on the people who use Web technology to create content - people like non-technical content authors, Web site creators, tool vendors, etc. These people don't need 45 pages of rules to reconcile conflicts between alt, <figure> and ARIA. These people need accessibility to be so simple that it blends seamlessly into normal work on the Web. They don't need us to create problems where there aren't any. And writing alternate text is not a problem. Simply read the text that replaces the image in the content that surrounds it, and if the text fits into the flow of surrounding content, you have perfect alternate text."

Again, in a world of unicorns and rainbows, your magic solution sounds great. But in reality, you have image that are informative, images that are decorative, images that are better represented by their purpose, images that contain text (not all of which may be relevant) and many other possible cases.

People like non-technical content authors, Web site creators, tool vendors, etc. do not need to understand all this. Why would they even want to in the first place? I'd need programing and design to be simpler too, but they're not. Why would accessibility be any different? It's a complex field with lots of grey zones and that requires some thinking, experience, depth. Oversimplifying like you do here not only doesn't help move forward, it actually pulls us all back because you imply it can be easy while years have shown us that it's not.

And anyone following your perfect rule would probably do better than most clueless authors out there are doing today, they would not score a perfect game because your rule simply isn't.

17. Posted by mgm
on Thursday 2011-01-20 at 00:55:01 PST

It seems that ALA also thinks that alt is only for the blind: "those who cannot see it".

>Accessibility is something that needs to be fought for, not begged for.
Right on! I feel that many accessibility experts have given up the fight and are now very cozy with people of influence and companies that just don't get accessibility. When it comes to accessibility, word "consensus" mean "sell out".

18. Posted by Hans Hillen
on Thursday 2011-01-20 at 03:32:19 PST

Alternative text is often considered to be the most basic, easiest to implement accessibility feature. You would be surprised though how many people and companies manage to screw it up. I've been doing accessibility testing and training for years and te situation is not good. People claim to know what they're doing when they speak about alt "tags" (which indeed they are often aware of through their WYSIWYG editor), but the alt attribute is misused more often than not. It either does not provide sufficient information (lack of an alt attribute, alt="image", alt="sys238723XXSHDJE", etc.), or is overdone (alt="Upper right rounded corner. This image is used for presentational purposes and should not be read by assistive technology").

This problem is even more apparent in corporate web applications that use more complex, automatically generated interfaces. In this environment images are used in a more complicated way than just a picture of president Obama, they're used as part of the application's UI structure. Figuring out proper alternative text in such a situation can be tricky, and your "less is more" approach is not going to cut it for people who don't have sufficient experience with a11y. Often such corporate applications are either created or tested by some offshore team in a far away country, by people who want to be instructed to the letter on how that task should be performed. The style guides and instructions used for this would really benefit from a good resource that explains exactly how alternative text can be applied (as the document indicates, there is more to it than just the alt attribute), and how to determine a proper text alternative. They need a complete, exhaustive overview. The W3C document currently being worked on is exactly that.

Did you learn xhtml by reading the W3C specs from beginning to end? I sure didn't. You learn through less dry resources such as tutorials, blog posts and books. But the specs are there as a reference. It's the same with this alt text document. There are many resources out there covering best practices for alt text, and they're all a little different. Now there is going to be an official, detailed resource that can serve as a base for the others. Whether or not it's 45 pages long is completely irrelevant. It's not meant for quick and easy reading, it's meant to cover everything there is to know about this topic. It's there for those that need it. People who don't need this level of detail can refer to smaller, more casual resources that do a good job explaining the most important rules (www.webAIM.org comes to mind as a great, reader friendly resource).

Of course, there is no reason why this lengthly W3C document couldn't have an accompanying quickref document, that just gives you the boiled down version of the material.

19. Posted by mattur
on Thursday 2011-01-20 at 07:37:54 PST

Vlad >> Here are stats that show how skewed the bug submission processed is against accessiblity issues:
Vlad > I did not offer any reason for this. However, I would be interested in knowing what you believe the reason is for the skewed stats.

All the charts show are the percentage breakdowns of the bug resolutions. To assess whether those resolutions are appropriate you would have to qualitatively assess the bugs.

Extreme example: starting from 0 bugs, I submit 1 typo bug, and 1 a11y-tf bug stating the spec should mandate the messages displayed by conformance tools. But the spec doesn't and cannot mandate any text for conformance tools. So the typo is FIXED and the a11y-tf bug is marked WONTFIX.

The charts would show main-WG bugs 100% FIXED and a11y-tf bugs 100% WONTFIX. Does this "show" the bug process is "skewed" against accessibility issues?

The latest version of the stats is here:
http://www.d.umn.edu/~lcarlson/html5bugchart/20101204/

With a bit more analysis here:
http://www.d.umn.edu/~lcarlson/html5bugchart/20101204/ld/chart1.html

20. Posted by mattur
on Thursday 2011-01-20 at 07:44:03 PST

> "Accessibility issues that affect everyone on the Web should be discussed and debated in public with full involvement of the accessibility community."

Agreed, and this is imho the reason why W3C WAI-PF group is systemically dysfunctional and unfit for purpose.

But SteveF has taken the lead here with this text alternatives doc, showing that the W3C /can/ produce reliable accessibility advice in a timely, open and accountable manner, using language that normal web authors can understand.

All W3C accessibility advice should be developed in this way.

21. Posted by Rachael
on Thursday 2011-01-20 at 08:57:34 PST

I think your points are good ones. So it's not to belittle them that I say, "Yes, I will read 45 pages about alt text."

Because for someone like me, who has no personal understanding of accessibility challenges, once you get into actually trying to write alt text for a specific image, you sometimes want confirmation & guidance beyond basic principles.

Well, if "you" is "me." I've probably already read 45 pages on alt text and image replacement and I'm excited about being about to read 45 more.

But I'm told I'm abnormal in my desire for long, detailed documentation and that I need to scale my own back a bit.

Sigh. No one at work reads my painstakingly put-together e-mails, either. Which is too bad, because if they did they wouldn't have to ask half the questions they ask for the next two weeks.

So, basically, while I think you're quite accurate, I also think there is a place for this document and web designers & developers who will appreciate it. Like me. Or at least I anticipate that I will. Guess that's not a guarantee.

22. Posted by Trev
on Thursday 2011-01-20 at 10:26:44 PST

I started reading this document and got maybe five pages into it before I was totally confused. I would like to know if Dennis, Mattur, Hans, Rachael have read this document in its +entirety+? If not, please do and let us know if you feel the same after reading the entire document.

23. Posted by Rachael
on Thursday 2011-01-20 at 10:53:51 PST

@Trev

I haven't read it at all yet. I just don't automatically think the *length* is a problem.

I said, "Or at least I anticipate that I will. Guess that's not a guarantee."

By which I meant, "It actually might just be a horribly obtuse thing to read and may not be the least bit useful, I don't know yet. However I am not put off by the length alone."

I will certainly judge it on its own merits once I have the opportunity to read it.

Why'd I comment at all if I hadn't? I may have missed this intimation, but I did not get the impression that Vlad found it confusing per se. The OP asked, "Will you read 45 pages on alt text?" He criticized its length and dryness.

I can & did say that I am not put off by the length, especially after glancing at the thumbnails which show plenty of illustrations. I can & did say that I feel there is enough complexity in real-life scenarios that 45 pages of information might be helpful. I can & did say, "Yeah, I'll read it."

I cannot yet say that this particular document is a success and have not yet done so. :-) Hope that clears up my position.

24. Posted by Joe Clark
on Friday 2011-01-21 at 05:58:35 PST

There *is* no W3C “consensus” process. They can simply take a vote to ignore your objections.

25. Posted by Jared Smith
on Friday 2011-01-21 at 07:33:26 PST

This has been an interesting discussion (despite the W3C process tangents). The truth of it is that alternative text is not and never will be a simple thing. Our site has two articles spanning 22 pages that cover just the basics (see http://webaim.org/techniques/alttext/). In our on-site trainings, we typically spend at least 90 minutes covering the nuances of alternative text - about the amount of time it would take to read 45 pages on the subject.

At it's very nature, alternative text is basic - content and function of the image. In real-life implementation, it is perhaps the most difficult aspect of modern web accessibility to truly understand. And this is not even to mention the added complexities of HTML5 (or whatever it's being called today) with optional alt and figcaption, or ARIA with aria-describedby and aria-labelledby.

If you want to really 'get' alternative text and implement it right, 45 pages probably isn't nearly enough.

27. Posted by BrianB
on Friday 2011-01-21 at 09:47:21 PST

Vlad, I don't think you undertand something. These so called accessibility experts don't want accessibility to be simple. They make money through tutorials, seminars, conferences, traning, books, etc and the more complexity, the more opportunity to sell services to the few who care about accessibility.

28. Posted by John foliot
on Friday 2011-01-21 at 14:29:05 PST

@BrianB: I today am fortunate to make a modest living as one of those so-called accessibility experts, but I struggled for nearly 5 years on below-poverty income learning and applying my craft and avocation. If you think that the hills of web-accessibility are paved with gold, lemme tell you how very wrong you are. Of the multitude of people I have had the pleasure to meet as "accessibility specialists" I can tell you, none of us are rich. Further, more often than not those "(online) tutorials, seminars, (and) conference" speaking engagements are done and offered for free: if you are interested in web accessibility, you should investigate either CSUN (registration required - http://www.csunconference.org/) or any number of the free-to-attend a11y un-conferences that are held in various cities throughout North America. AFAIK, none of the speakers at those events are ever paid, and admission is usually free. Of the numerous public speaking engagements I have given over the years, I have received in total exactly 1 honorarium of $300.00, so I am not yet ready to retire on that. However, I and others (just like you) can also be contractually engaged for professional services, at rates mutually agreed upon - I mean, you too do work for a living, right?

@Vlad: Vlad, Vlad, Vlad... once again you look at a problem with your curiously rose-colored glasses and pronounce a 'solution' that has very little to do with real life. Achieving real web accessibility is complex - pretending that it can be made magically 'easy' by using Vlad's Simple Rules of Easy Accessibility does web accessibility and those affected by it such a dis-service I feel compelled to continually enter into debate with you. You are wrong. It is as simple as that.

If content authors can't or won't bother to read a detailed document, or attend seminars and sessions created by specialists who *have* read and understand the complex and subtle nuances of creating appropriate alternative text, then there is very little we can do about it. There is no silver bullet or magical incantation that will eliminate the hard work required to achieve success. Is this fair or beneficial to disabled users? Not at all. Is it a reality and truth we must accept? Sadly yes. As blues musician Albert King once wrote: "Everybody wants to go to heaven, nobody wants to die".

You can argue until the end of time that creating appropriate alt text is easy, and you will still and always be wrong: there is more to driving a car than learning how to start the motor and engage the transmission, and suggesting anything else simply proves that *you* don't fully understand the problems of web accessibility in the real world - a point I have long ago concluded.