WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: Investigating the proposed alt attribute recommendations in HTML 5

for

Number of posts in this thread: 27 (In chronological order)

From: Steve Faulkner
Date: Thu, Aug 30 2007 2:50AM
Subject: Investigating the proposed alt attribute recommendations in HTML 5
No previous message | Next message →

If the developers of flickr.com or Photobucket were to implement the
recommendations regarding the omission of the alt attribute within the
current HTML 5 draft what are the potential effects upon the accessibility
of the sites for users of assistive technology such as screen readers?

Investigating the proposed alt attribute recommendations in HTML 5 -
http://www.paciellogroup.com/resources/articles/altinhtml5.html



--
with regards

Steve Faulkner
Technical Director - TPG Europe
Director - Web Accessibility Tools Consortium

www.paciellogroup.com | www.wat-c.org
Web Accessibility Toolbar -
http://www.paciellogroup.com/resources/wat-ie-about.html

From: Christian Heilmann
Date: Thu, Aug 30 2007 3:10AM
Subject: Re: Investigating the proposed alt attribute recommendations in HTML 5
← Previous message | Next message →

> If the developers of flickr.com or Photobucket were to implement the
> recommendations regarding the omission of the alt attribute within the
> current HTML 5 draft what are the potential effects upon the accessibility
> of the sites for users of assistive technology such as screen readers?
>
> Investigating the proposed alt attribute recommendations in HTML 5 -
> http://www.paciellogroup.com/resources/articles/altinhtml5.html

The easier question is what is the impact on SEO, as this is where a
money loss would be felt which sadly enough makes people listen more
eagerly.

--
Chris Heilmann
Book: http://www.beginningjavascript.com
Blog: http://www.wait-till-i.com
Writing: http://icant.co.uk/

From: Austin, Darrel
Date: Thu, Aug 30 2007 8:00AM
Subject: Re: Investigating the proposed alt attribute recommendationsin HTML 5
← Previous message | Next message →

> Investigating the proposed alt attribute recommendations in HTML 5 -
> http://www.paciellogroup.com/resources/articles/altinhtml5.html

Seems like a change that Target Corp. lobbied for. ;o)

-Darrel

From: Stephanie Sullivan
Date: Thu, Aug 30 2007 8:10AM
Subject: Re: Investigating the proposed alt attribute recommendations in HTML 5
← Previous message | Next message →

On Aug 30, 2007, at 5:06 AM, Christian Heilmann wrote:

> The easier question is what is the impact on SEO, as this is where a
> money loss would be felt which sadly enough makes people listen more
> eagerly.

I've come to believe that the alt attribute doesn't give a very large
SEO bump these days. I believe this is due to the blatant abuse and
stuffing of these attributes with large amounts of keywords. I can
definitely confirm by a site I was hired to correct that Google seems
to ignore very verbose alts (they had used images instead of text for
much of the site. Then, realizing their error, they tried to put
whole sentences, and sometimes paragraphs of text, into the alt
attribute to make up for it). The words that were in their alts were
not in Google's database.

That said, I can not confirm or deny whether small, descriptive alts
are still considered in any significant way in the algorithm to
target a specific page or for their image search capabilities. (Good
thinking though Christian and likely something that should be looked
at since sadly, money talks louder than accessibility.)

(Thanks for the AT information, Steve. :) I noticed to that Flickr
uses the repetitive alt in the pages with multiple pictures, but on
single picture pages uses an empty alt.)

---
Stephanie Sullivan
http://www.w3conversions.com
Dreamweaver Task Force for WaSP
http://www.communitymx.com




From: Andrew Kirkpatrick
Date: Thu, Aug 30 2007 8:40AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

Maybe I'm misunderstanding the situation, but HTML5 is not removing the
alt attribute, just not making it required. For linked images, omitting
the alt attribute and not having a good value would still be an
accessiblity issue, and it would be valid and expected to add alt="Boba
and Franny's Baby Cat Sleeping". I think that your page highlights why
it is necessary to use the alt attribute on images, but it doesn't
address what the downside of not requiring the alt attribute for all
images.

I feel like this change is being received as a stab in the back to all
the good work in accessibility that has taken place over the years, but
I don't think that is the case....

AWK


> -----Original Message-----
> From: = EMAIL ADDRESS REMOVED =
> [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of
> Christian Heilmann
> Sent: Thursday, August 30, 2007 5:07 AM
> To: WebAIM Discussion List
> Subject: Re: [WebAIM] Investigating the proposed alt
> attributerecommendations in HTML 5
>
> > If the developers of flickr.com or Photobucket were to
> implement the
> > recommendations regarding the omission of the alt attribute
> within the
> > current HTML 5 draft what are the potential effects upon the
> > accessibility of the sites for users of assistive
> technology such as screen readers?
> >
> > Investigating the proposed alt attribute recommendations in
> HTML 5 -
> > http://www.paciellogroup.com/resources/articles/altinhtml5.html
>
> The easier question is what is the impact on SEO, as this is
> where a money loss would be felt which sadly enough makes
> people listen more eagerly.
>
> --
> Chris Heilmann
> Book: http://www.beginningjavascript.com
> Blog: http://www.wait-till-i.com
> Writing: http://icant.co.uk/
>

From: Stephanie Sullivan
Date: Thu, Aug 30 2007 9:00AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

On Aug 30, 2007, at 10:30 AM, Andrew Kirkpatrick wrote:

> I think that your page highlights why
> it is necessary to use the alt attribute on images, but it doesn't
> address what the downside of not requiring the alt attribute for all
> images.

Can you expound on that Andrew?

> I feel like this change is being received as a stab in the back to all
> the good work in accessibility that has taken place over the years,
> but
> I don't think that is the case....

To me, it puts it in the realm of that "relying on the user to
understand and make the right choice" problem which exists in much of
the accessibility world. Those of us that do lots of web support and
send people to the validator to sort out problems, before we spend
time dissecting what odd coding issue they've created, know the value
of having the alt required. Many of these newer developers (or maybe
they're not new, they've just never run a page through the validator
and don't keep up with our business) haven't a clue that they should
use an alt attribute. Yes, here at WebAIM, it seems ludicrous. But I
see it every day. Every. Day.

To me, the benefits of having it required far outweigh the possible
downside in a few unique situations. You're now leaving it up to
developers to read the specs and know what things might be a good
idea. They don't do it. Thus, we lose an accessibility hook that in
nearly all cases is the best option. Maybe I'm misunderstanding what
cases exist where the alt becomes a negative attribute?

---
Stephanie Sullivan
http://www.w3conversions.com
Dreamweaver Task Force for WaSP
http://www.communitymx.com




From: Tim Beadle
Date: Thu, Aug 30 2007 9:40AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

On 30/08/2007, Stephanie Sullivan < = EMAIL ADDRESS REMOVED = > wrote:
> Maybe I'm misunderstanding what
> cases exist where the alt becomes a negative attribute?

When keen but ill-informed developers add alt="This is the ACME, Inc.
logo, a picture of a Stork" instead of just alt="ACME, Inc."? Or they
repeat information about the image that is nearby in the text anyway.

Having alt as a required attribute doesn't in any way guarantee
quality, useful alt text.

Tim

From: Gez Lemon
Date: Thu, Aug 30 2007 9:50AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

Hi Tim,

On 30/08/2007, Tim Beadle < = EMAIL ADDRESS REMOVED = > wrote:
> On 30/08/2007, Stephanie Sullivan < = EMAIL ADDRESS REMOVED = > wrote:
> > Maybe I'm misunderstanding what
> > cases exist where the alt becomes a negative attribute?
>
> When keen but ill-informed developers add alt="This is the ACME, Inc.
> logo, a picture of a Stork" instead of just alt="ACME, Inc."? Or they
> repeat information about the image that is nearby in the text anyway.
>
> Having alt as a required attribute doesn't in any way guarantee
> quality, useful alt text.

It doesn't guarantee quality useful alternate text, but the answer is
not to make alternate text optional. The real issue is ensuring that
tool vendors should make it easier for their users to provide
alternate text, and educate users as to what makes good alternate
text. Making the alt attribute doesn't solve either of those issues.


Gez


--

From: Debi Orton
Date: Thu, Aug 30 2007 10:00AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

At 10:58 AM 8/30/2007, Stephanie Sullivan wrote:

><snip>

>To me, it puts it in the realm of that "relying on the user to
>understand and make the right choice" problem which exists in much of
>the accessibility world. Those of us that do lots of web support and
>send people to the validator to sort out problems, before we spend
>time dissecting what odd coding issue they've created, know the value
>of having the alt required. Many of these newer developers (or maybe
>they're not new, they've just never run a page through the validator
>and don't keep up with our business) haven't a clue that they should
>use an alt attribute. Yes, here at WebAIM, it seems ludicrous. But I
>see it every day. Every. Day.
>
>To me, the benefits of having it required far outweigh the possible
>downside in a few unique situations. You're now leaving it up to
>developers to read the specs and know what things might be a good
>idea. They don't do it. Thus, we lose an accessibility hook that in
>nearly all cases is the best option. Maybe I'm misunderstanding what
>cases exist where the alt becomes a negative attribute?
>
>---
>Stephanie Sullivan
>http://www.w3conversions.com
>Dreamweaver Task Force for WaSP
>http://www.communitymx.com
>



Not to mention all of the newly-minted web developers whose only
training was in application development, where the words 'usability'
and 'accessibility' have not yet entered the lexicon.

Here's one personal example. Last spring two co-workers and I
attended a week-long XML class. Out of 13 students, we three were
the only ones who had ever coded a web page.

When we got to the XSLT portion of the class, the instructor -- a
very knowledgable man with regard to XML as a programming tool --
provided code examples that would have choked a validator (e.g.,
'FONT,' CENTER,' omission of closing tags -- all this in a document
with an XHTML doctype).

I pointed out that NYS had a policy with regard to the accessibility
of web content, and his reaction was "That's nice, but what does it
have to do with XML?" Surely NY government can't be the only
organization in which this migration is occurring.

Unfortunately, I'm afraid that REQUIRING accommodations is the only
way to ensure accessible content.

Debi Orton


From: Steven Faulkner
Date: Thu, Aug 30 2007 10:10AM
Subject: Re: Investigating the proposed alt attributerecommendations in HTML 5
← Previous message | Next message →

Hi Andrew,

>Maybe I'm misunderstanding the situation, but HTML5 is not removing the
>alt attribute, just not making it required.

No you are not misunderstanding the situation. What I am arguing is
reasoning given for that the change of the alt attribute to optional and
recommending its omission (as it does in the spec) is not based on any sound
evidence and if a little research and testing it is done it is found that
(for screen reader users for example) it has the potential to have an
adverse effect upon accessibility. Therefore why not leave the alt as
required?

Some "real world" use cases need to be provided where the omission of the
alt is beneficial, as yet they have not been.


On 30/08/2007, Andrew Kirkpatrick < = EMAIL ADDRESS REMOVED = > wrote:
>
> Maybe I'm misunderstanding the situation, but HTML5 is not removing the
> alt attribute, just not making it required. For linked images, omitting
> the alt attribute and not having a good value would still be an
> accessiblity issue, and it would be valid and expected to add alt="Boba
> and Franny's Baby Cat Sleeping". I think that your page highlights why
> it is necessary to use the alt attribute on images, but it doesn't
> address what the downside of not requiring the alt attribute for all
> images.
>
> I feel like this change is being received as a stab in the back to all
> the good work in accessibility that has taken place over the years, but
> I don't think that is the case....
>
> AWK
>
>
> > -----Original Message-----
> > From: = EMAIL ADDRESS REMOVED =
> > [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of
> > Christian Heilmann
> > Sent: Thursday, August 30, 2007 5:07 AM
> > To: WebAIM Discussion List
> > Subject: Re: [WebAIM] Investigating the proposed alt
> > attributerecommendations in HTML 5
> >
> > > If the developers of flickr.com or Photobucket were to
> > implement the
> > > recommendations regarding the omission of the alt attribute
> > within the
> > > current HTML 5 draft what are the potential effects upon the
> > > accessibility of the sites for users of assistive
> > technology such as screen readers?
> > >
> > > Investigating the proposed alt attribute recommendations in
> > HTML 5 -
> > > http://www.paciellogroup.com/resources/articles/altinhtml5.html
> >
> > The easier question is what is the impact on SEO, as this is
> > where a money loss would be felt which sadly enough makes
> > people listen more eagerly.
> >
> > --
> > Chris Heilmann
> > Book: http://www.beginningjavascript.com
> > Blog: http://www.wait-till-i.com
> > Writing: http://icant.co.uk/
> >

From: Andrew Kirkpatrick
Date: Thu, Aug 30 2007 10:20AM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

> > I think that your page highlights why
> > it is necessary to use the alt attribute on images, but it doesn't
> > address what the downside of not requiring the alt
> attribute for all
> > images.
>
> Can you expound on that Andrew?

The basic idea that is being advocated in HTML5 is that alt should be
used when it is needed. If alt has a null value, then it isn't really
needed, so could be omitted. There are lots of attributes that function
in this way, and I'm arguing that since there isn't additional value in
requiring alt for images that don't need it, alt can be treated this way
without adverse affects.

Steve's page is mostly highlighting what happens if you don't use the
alt attribute when it is needed, and there are known problems for users.
This isn't really the point, since alt is still allowed on images.

Requiring alt on all images would mean that pages (assuming that they
are valid) containing images would have alt attributes, but nothing to
ensure that the alt attributes have values or useful values.

> To me, it puts it in the realm of that "relying on the user
> to understand and make the right choice" problem which exists
> in much of the accessibility world. Those of us that do lots

Do you think that we're not not now in that realm anyway?

> ludicrous. But I see it every day. Every. Day.

I'm not exactly cloistered away on the WebAIM list all day... :)

I don't have a real objection to alt being required or not. All I want
to point out is that there is nothing inherently wrong about having it
not be. <img src="spacer.gif" alt="" /> is identical to <img
src="spacer.gif" /> semantically. Either way, Dreamweaver will still
prompt for alt when you insert an image, but you are correct that people
who only find out about accessibility is by validating their site will
lose an opportunity to educate themseleves. On the other hand, we won't
have any one doing a find and replace for '<img' and replacing it with
'<img alt=""' to get by the validator. If a page has no alt values,
then it is laid bare that they haven't attended to accessibility (unless
they ONLY use spacer images).

Images are probably less of an issue than form labels. Should we
require the id attribute on inputs and the for attribute on label
elements? Require the caption element inside tables? People developing
for the web need to learn how to code properly or use tools that
encourage it, but I just don't think that making alt not required is a
problem.

AWK


From: Vlad Alexander (XStandard)
Date: Thu, Aug 30 2007 10:40AM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

Andrew wrote:

> <img src="spacer.gif" alt="" /> is identical to <img
> src="spacer.gif" /> semantically.
That may be true for decorative images. But it certainly is not true for non-decorative images such as photos or diagrams. The HTML 5 spec _is_ advocating making the the alt text optional for non-decorative images such as photos. So it is condoning the practice of making Web content that is not accessible by people who cannot see. That is wrong. The Web should be accessible to everyone, not only for some groups of people.

Regards,
-Vlad
http://xhtml.com


-------- Original Message --------
From: Andrew Kirkpatrick
Date: 2007-08-30 12:14 PM
>>> I think that your page highlights why
>>> it is necessary to use the alt attribute on images, but it doesn't
>>> address what the downside of not requiring the alt
>> attribute for all
>>> images.
>> Can you expound on that Andrew?
>
> The basic idea that is being advocated in HTML5 is that alt should be
> used when it is needed. If alt has a null value, then it isn't really
> needed, so could be omitted. There are lots of attributes that function
> in this way, and I'm arguing that since there isn't additional value in
> requiring alt for images that don't need it, alt can be treated this way
> without adverse affects.
>
> Steve's page is mostly highlighting what happens if you don't use the
> alt attribute when it is needed, and there are known problems for users.
> This isn't really the point, since alt is still allowed on images.
>
> Requiring alt on all images would mean that pages (assuming that they
> are valid) containing images would have alt attributes, but nothing to
> ensure that the alt attributes have values or useful values.
>
>> To me, it puts it in the realm of that "relying on the user
>> to understand and make the right choice" problem which exists
>> in much of the accessibility world. Those of us that do lots
>
> Do you think that we're not not now in that realm anyway?
>
>> ludicrous. But I see it every day. Every. Day.
>
> I'm not exactly cloistered away on the WebAIM list all day... :)
>
> I don't have a real objection to alt being required or not. All I want
> to point out is that there is nothing inherently wrong about having it
> not be. <img src="spacer.gif" alt="" /> is identical to <img
> src="spacer.gif" /> semantically. Either way, Dreamweaver will still
> prompt for alt when you insert an image, but you are correct that people
> who only find out about accessibility is by validating their site will
> lose an opportunity to educate themseleves. On the other hand, we won't
> have any one doing a find and replace for '<img' and replacing it with
> '<img alt=""' to get by the validator. If a page has no alt values,
> then it is laid bare that they haven't attended to accessibility (unless
> they ONLY use spacer images).
>
> Images are probably less of an issue than form labels. Should we
> require the id attribute on inputs and the for attribute on label
> elements? Require the caption element inside tables? People developing
> for the web need to learn how to code properly or use tools that
> encourage it, but I just don't think that making alt not required is a
> problem.
>
> AWK
>
>
>

From: James Leslie
Date: Thu, Aug 30 2007 10:50AM
Subject: Re: Investigating the proposed altattributerecommendationsin HTML 5
← Previous message | Next message →

How are user agents supposed to distinguish between alt text that has
intentionally been omitted, and alt text that has been omitted because
the user didn't understand what was required of them, or their tool made
it difficult for them to provide the information? At the moment, alt=""
lets the user agent know that they don't need to use repair techniques
to provide the missing information to the user, provided it isn't used
in a critical way (such as being the sole way of exposing a link
phrase). If the attribute becomes optional, people who care about
standards but not about accessibility won't provide the alt attribute.
That is definitely a step backwards.

---------------

It seems a step backwards to me to. Surely decorative images should be
placed on the page using CSS, meaning that the alt tag is not
required/used anyway?

The argument mainly seems to be that a bunch of 'developers/designers'
don't know how to use the alt tag properly so instead of trying to
educate them, we deprecate a useful attribute. Either that or we are
allowing big commercial sites like Flickr to decide how standards should
be deployed, in order to fit in with their model. It seems like a
dangerous path to be taking to me.

From: Gez Lemon
Date: Thu, Aug 30 2007 11:00AM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

Hi Andrew,

> The basic idea that is being advocated in HTML5 is that alt should be
> used when it is needed. If alt has a null value, then it isn't really
> needed, so could be omitted.

How are user agents supposed to distinguish between alt text that has
intentionally been omitted, and alt text that has been omitted because
the user didn't understand what was required of them, or their tool
made it difficult for them to provide the information? At the moment,
alt="" lets the user agent know that they don't need to use repair
techniques to provide the missing information to the user, provided it
isn't used in a critical way (such as being the sole way of exposing a
link phrase). If the attribute becomes optional, people who care about
standards but not about accessibility won't provide the alt attribute.
That is definitely a step backwards.


Gez


--

From: Steve Faulkner
Date: Thu, Aug 30 2007 11:10AM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

Hi Andrew,

>Steve's page is mostly highlighting what happens if you don't use the
>alt attribute when it is needed, and there are known problems for users.
>This isn't really the point, since alt is still allowed on images.

yes, but these are the sites cited as containing images that should have the
alt omitted.

>All I want
>to point out is that there is nothing inherently wrong about having it
>not be. <img src="spacer.gif" alt="" /> is identical to <img
>src="spacer.gif" /> semantically. Either way

This is not what the current version of the spec is advocating, in the case
of spacer images the spec states:

"The img<http://www.whatwg.org/specs/web-apps/current-work/multipage/section-embedded.html#img>;must
not be used as a layout tool. In particular,
img<http://www.whatwg.org/specs/web-apps/current-work/multipage/section-embedded.html#img>;elements
should not be used to display fully transparent images, as they
rarely convey meaning and rarely add anything useful to the document. "

In the case of decorative images its states:

"A purely decorative image that doesn't add any information but is still
specific to the surrounding content

In some cases, the image isn't discussed by the surrounding text, but it has
some relevance. Such images are decorative, but still form part of the
content. In these cases, *the
alt<http://www.whatwg.org/specs/web-apps/current-work/multipage/section-embedded.html#alt>;attribute
must be present but its value must be the empty string
*."
in regards to the omission of the alt:

"*When the alt<http://www.whatwg.org/specs/web-apps/current-work/multipage/section-embedded.html#alt>;attribute
is missing, the image represents a key part of the content
*. Non-visual user agents should apply image analysis heuristics to help the
user make sense of the image."


HTML 5 draft specification [
http://www.whatwg.org/specs/web-apps/current-work/multipage/section-embedded.html#the-img
]

On 30/08/2007, Andrew Kirkpatrick < = EMAIL ADDRESS REMOVED = > wrote:
>
> > > I think that your page highlights why
> > > it is necessary to use the alt attribute on images, but it doesn't
> > > address what the downside of not requiring the alt
> > attribute for all
> > > images.
> >
> > Can you expound on that Andrew?
>
> The basic idea that is being advocated in HTML5 is that alt should be
> used when it is needed. If alt has a null value, then it isn't really
> needed, so could be omitted. There are lots of attributes that function
> in this way, and I'm arguing that since there isn't additional value in
> requiring alt for images that don't need it, alt can be treated this way
> without adverse affects.
>
> Steve's page is mostly highlighting what happens if you don't use the
> alt attribute when it is needed, and there are known problems for users.
> This isn't really the point, since alt is still allowed on images.
>
> Requiring alt on all images would mean that pages (assuming that they
> are valid) containing images would have alt attributes, but nothing to
> ensure that the alt attributes have values or useful values.
>
> > To me, it puts it in the realm of that "relying on the user
> > to understand and make the right choice" problem which exists
> > in much of the accessibility world. Those of us that do lots
>
> Do you think that we're not not now in that realm anyway?
>
> > ludicrous. But I see it every day. Every. Day.
>
> I'm not exactly cloistered away on the WebAIM list all day... :)
>
> I don't have a real objection to alt being required or not. All I want
> to point out is that there is nothing inherently wrong about having it
> not be. <img src="spacer.gif" alt="" /> is identical to <img
> src="spacer.gif" /> semantically. Either way, Dreamweaver will still
> prompt for alt when you insert an image, but you are correct that people
> who only find out about accessibility is by validating their site will
> lose an opportunity to educate themseleves. On the other hand, we won't
> have any one doing a find and replace for '<img' and replacing it with
> '<img alt=""' to get by the validator. If a page has no alt values,
> then it is laid bare that they haven't attended to accessibility (unless
> they ONLY use spacer images).
>
> Images are probably less of an issue than form labels. Should we
> require the id attribute on inputs and the for attribute on label
> elements? Require the caption element inside tables? People developing
> for the web need to learn how to code properly or use tools that
> encourage it, but I just don't think that making alt not required is a
> problem.
>
> AWK
>
>
>

From: Karl Groves
Date: Thu, Aug 30 2007 11:20AM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | Next message →

I really don't think it matters whether the alt attribute is required or
not. Truth is, incomplete/ inaccurate/ deceptive alt text is just as bad as
none at all. It seems as though some people believe that requiring it will
somehow mean people will use it (and use it effectively). That's not the
case at all! As Tim Beadle (I think) said, there's a lot of really BAD alt
text out there and the alt attribute is required now. Clearly, requiring it
hasn't meant a thing to people who either don't care or don't know about
supplying good alt text.

None of this means that I agree with making it optional, but rather that the
hysteria is unfounded. Requiring it for validity's sake doesn't mean that,
by proxy, websites suddenly become more accessible.

Karl

From: Steve Faulkner
Date: Thu, Aug 30 2007 11:30AM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | Next message →

>incomplete/ inaccurate/ deceptive alt text is just as bad as none at all

I beg to differ, in many cases even poor quality alt text is preferable for
AT users as it may provide some information about the image.

>but rather that the hysteria is unfounded.

what behaviour has been demonstrated by anyone that could be described as
"hysteria"?

On 30/08/2007, Karl Groves < = EMAIL ADDRESS REMOVED = > wrote:
>
>
> I really don't think it matters whether the alt attribute is required or
> not. Truth is, incomplete/ inaccurate/ deceptive alt text is just as bad
> as
> none at all. It seems as though some people believe that requiring it
> will
> somehow mean people will use it (and use it effectively). That's not the
> case at all! As Tim Beadle (I think) said, there's a lot of really BAD alt
> text out there and the alt attribute is required now. Clearly, requiring
> it
> hasn't meant a thing to people who either don't care or don't know about
> supplying good alt text.
>
> None of this means that I agree with making it optional, but rather that
> the
> hysteria is unfounded. Requiring it for validity's sake doesn't mean that,
> by proxy, websites suddenly become more accessible.
>
> Karl
>
>

From: Jukka K. Korpela
Date: Thu, Aug 30 2007 11:40AM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | Next message →

On Thu, 30 Aug 2007, Karl Groves wrote:

> I really don't think it matters whether the alt attribute is required or
> not.

Making it not required would be a strong signal.

> Truth is, incomplete/ inaccurate/ deceptive alt text is just as bad as
> none at all.

Sometimes worse. That doesn't mean that omitting an alt attribute would
ever be acceptable. There has been too much talk about using alt
attributes as compared with emphasizing how they should be written, but
the solution of making them not required is just crazy.

Even _considering_ the issue is a sure symptom of the "HTML 5" activity
being on the wrong track. The whole "project" really puzzles me.

--
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/

From: Gez Lemon
Date: Thu, Aug 30 2007 11:50AM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | Next message →

Hi Karl,

On 30/08/2007, Karl Groves < = EMAIL ADDRESS REMOVED = > wrote:
> I really don't think it matters whether the alt attribute is required or
> not. Truth is, incomplete/ inaccurate/ deceptive alt text is just as bad as
> none at all. It seems as though some people believe that requiring it will
> somehow mean people will use it (and use it effectively). That's not the
> case at all! As Tim Beadle (I think) said, there's a lot of really BAD alt
> text out there and the alt attribute is required now. Clearly, requiring it
> hasn't meant a thing to people who either don't care or don't know about
> supplying good alt text.
>
> None of this means that I agree with making it optional, but rather that the
> hysteria is unfounded. Requiring it for validity's sake doesn't mean that,
> by proxy, websites suddenly become more accessible.

Hysteria is a bit strong, as most people are just responding to
something that is up for discussion. It would be a very sad state of
affairs if people were afraid to speak up about drafts through fear of
being told they're hysterical, particularly as drafts rely on feedback
from interested stakeholders.

I don't think that those commenting on this issue are naive enough to
believe that just because an attribute is required that it will be
used correctly, but are concerned that by making it not required is a
step backwards. Structurally, alternate text is an important part of
an image. As a markup language should be concerned with structure, it
should be reason enough to insist that the alt attribute is a required
attribute. The suggestion to make it optional has been made purely on
the basis that there are applications that generate poor alternate
text. The real solution is to get those applications to help users
provide good alternate text, not to make the attribute optional, as
that isn't helping anyone.

It's incredible to think that in 2007 people are explaining why
alternate text is important on a web accessibility mailing list.


Gez


--

From: Karl Groves
Date: Thu, Aug 30 2007 12:00PM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | Next message →

>>incomplete/ inaccurate/ deceptive alt text is just as bad as none at all

> I beg to differ, in many cases even poor quality alt text is preferable
for AT users as it may provide some information about the image.


A vast majority of my time is spent reviewing work performed by others on US
Government websites. In *many* instances, these "others" aren't exactly
what I'd call "experts" when it comes to accessibility. Nevertheless,
they're told that in order to adhere to Section 508 they must supply alt
attributes for all images. (In fact, they sometimes think alt text is all
that's required).

In most instances, the alt text supplied is, as I've said, incomplete,
inaccurate, or deceptive. Sometimes the text supplied isn't terrible and
just needs a little improvement. But in other cases, I've seen alt text so
bad that the user would be better off not having any at all. Unfortunately,
I'm bound by confidentiality from revealing the details of what I've seen,
but suffice it to say that the user would have been better off with null alt
text. Heck, some alt text I've seen has been so bad, having the screen
reader read the image name would have been more helpful.




>>but rather that the hysteria is unfounded.
> what behaviour has been demonstrated by anyone that could be described as
"hysteria"?

>From my reading of these messages, it would appear that others hold the view
that if the alt attribute is made optional, all of a sudden the entire web
will become inaccessible. I hold a differing view - that the web already
*is* inaccessible, requiring the alt attribute or not will make no
difference. The only way to make the web more accessible is by educating
content producers on how to do so.


Karl




On 30/08/2007, Karl Groves < = EMAIL ADDRESS REMOVED = > wrote:


I really don't think it matters whether the alt attribute is
required or
not. Truth is, incomplete/ inaccurate/ deceptive alt text
is just as bad as
none at all. It seems as though some people believe that
requiring it will
somehow mean people will use it (and use it effectively).
That's not the
case at all! As Tim Beadle (I think) said, there's a lot of
really BAD alt
text out there and the alt attribute is required now.
Clearly, requiring it
hasn't meant a thing to people who either don't care or
don't know about
supplying good alt text.

None of this means that I agree with making it optional, but
rather that the
hysteria is unfounded. Requiring it for validity's sake
doesn't mean that,
by proxy, websites suddenly become more accessible.

Karl

From: Christophe Strobbe
Date: Thu, Aug 30 2007 1:30PM
Subject: Re: Investigating the proposed alt attribute recommendations in HTML 5
← Previous message | Next message →

Hi,

At 19:23 30/08/2007, Steve Faulkner wrote:
>(...)
> >but rather that the hysteria is unfounded. [Karl Groves]
>
>what behaviour has been demonstrated by anyone that could be described as
>"hysteria"?


May I point out that the HTML 5 WG has "so far, made *no* design decisions"?
(Dan Connolly at
<http://lists.w3.org/Archives/Public/w3c-wai-ig/2007JulSep/0047.html>;).
There is currently only an editor's draft, no "Working Draft".

Best regards,

Christophe


--
Christophe Strobbe
K.U.Leuven - Dept. of Electrical Engineering - SCD
Research Group on Document Architectures
Kasteelpark Arenberg 10 bus 2442
B-3001 Leuven-Heverlee
BELGIUM
tel: +32 16 32 85 51
http://www.docarch.be/


Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm

From: Gez Lemon
Date: Thu, Aug 30 2007 1:40PM
Subject: Re: Investigating the proposed alt attribute recommendations in HTML 5
← Previous message | Next message →

Hi Christophe,

On 30/08/2007, Christophe Strobbe < = EMAIL ADDRESS REMOVED = > wrote:
> May I point out that the HTML 5 WG has "so far, made *no* design decisions"?
> (Dan Connolly at
> <http://lists.w3.org/Archives/Public/w3c-wai-ig/2007JulSep/0047.html>;).
> There is currently only an editor's draft, no "Working Draft".

Yes, of course. Some of the people commenting, including Steve
Faulkner, are members of the HTML 5 working group. As part of making
sensible design decisions within the group regarding accessibility,
they have decided to include the accessibility community. As I'm sure
you're aware, the earlier that accessibility is discussed within the
design phase, the better, which is why it is a good thing that the
accessibility advocates within the HTML working group have opened this
up to the wider community.

Best regards,

Gez



--

From: Steve Faulkner
Date: Thu, Aug 30 2007 2:10PM
Subject: Re: Investigating the proposed alt attribute recommendations in HTML 5
← Previous message | Next message →

christophe wrote:
>May I point out that the HTML 5 WG has "so far, made *no* design
decisions"?

that's why the title of the article i wrote is

"Investigating the PROPOSED alt attribute recommendations in HTML 5."


On 30/08/2007, Christophe Strobbe < = EMAIL ADDRESS REMOVED = >
wrote:
>
> Hi,
>
> At 19:23 30/08/2007, Steve Faulkner wrote:
> >(...)
> > >but rather that the hysteria is unfounded. [Karl Groves]
> >
> >what behaviour has been demonstrated by anyone that could be described as
> >"hysteria"?
>
>
> May I point out that the HTML 5 WG has "so far, made *no* design
> decisions"?
> (Dan Connolly at
> <http://lists.w3.org/Archives/Public/w3c-wai-ig/2007JulSep/0047.html>;).
> There is currently only an editor's draft, no "Working Draft".
>
> Best regards,
>
> Christophe
>
>
> --
> Christophe Strobbe
> K.U.Leuven - Dept. of Electrical Engineering - SCD
> Research Group on Document Architectures
> Kasteelpark Arenberg 10 bus 2442
> B-3001 Leuven-Heverlee
> BELGIUM
> tel: +32 16 32 85 51
> http://www.docarch.be/
>
>
> Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
>
>

From: Jared Smith
Date: Thu, Aug 30 2007 3:40PM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

On 8/30/07, John Foliot - Stanford Online Accessibility Program
< = EMAIL ADDRESS REMOVED = > wrote:

> The use-case(s) being proposed might seem semi-valid (Flickr et. al.)

The biggest problem with this whole thing is that there are a few
cases where missing alt attribute or an empty alt attribute will
result in more than just a bit of missing information. For instance,
when an image is the only thing inside a link, *it must have
descriptive alt text*. If it does not, the screen reader must present
something to represent the function of that link - typically it will
read the link URL, something which is rarely very useful. On sites
like Flickr, the images are almost always a link - removing the alt
text or making it null will be very detrimental for accessibility.

In short, the whole approach to this seems a bit backward. While
functionally there is little difference between null alt text and no
alt attribute at all, making the attribute optional is inviting
problems. At a minimum, the spec should require functional alt text
when the image is the only thing within an anchor tag as that is the
*ONLY* method of making such links accessible.

Jared Smith
WebAIM

From: John Foliot - Stanford Online Accessibility Program
Date: Thu, Aug 30 2007 3:50PM
Subject: Re: Investigating the proposed altattributerecommendations in HTML 5
← Previous message | Next message →

Tim Beadle wrote:
>
> When keen but ill-informed developers add alt="This is the ACME, Inc.
> logo, a picture of a Stork" instead of just alt="ACME, Inc."? Or they
> repeat information about the image that is nearby in the text anyway.
>
> Having alt as a required attribute doesn't in any way guarantee
> quality, useful alt text.

Granted. The issue is of course that they need to *think* about what they
are using for alt text. Removing the mandate to have alt text (by making it
optional) makes it easier for them to simply not use alt text (it's a
no-brainer) - and then point to the spec and say "see, it says optional".

The use-case(s) being proposed might seem semi-valid (Flickr et. al.), but
making the alt attribute optional opens the door for mis-use, and contrary
to what has been suggested, is an apparent step back from ground already
fought for.

Making alt text optional acknowledges the problem, but does nothing to fix
the problem except make it easier to ignore the problem.

JF



From: Andrew Kirkpatrick
Date: Thu, Aug 30 2007 4:50PM
Subject: Re: Investigating the proposed altattributerecommendationsin HTML 5
← Previous message | Next message →

> in regards to the omission of the alt:
>
> "*When the
> alt<http://www.whatwg.org/specs/web-apps/current-work/multipag
e/section-embedded.html#alt>attribute
> is missing, the image represents a key part of the content *.
> Non-visual user agents should apply image analysis heuristics
> to help the user make sense of the image."
>
>
> HTML 5 draft specification [
> http://www.whatwg.org/specs/web-apps/current-work/multipage/se
ction-embedded.html#the-img

OK, reading that section again, it is referring to images that have no
equivalent and "doesn't have an obviously textual alternative". The
spec says:

<snip>
In certain rare cases, the image is simply a critical part of the
content, and there might even be no alternative text available. This
could be the case, for instance, in a photo gallery, where a user has
uploaded 3000 photos from a vacation trip, without providing any
descriptions of the images. The images are the whole point of the pages
containing them.

In such cases, the alt attribute may be omitted, but the alt attribute
should be included, with a useful value, if at all possible. If an image
is a key part of the content, the alt attribute must not be specified
with an empty value.
</snip>

To me this is attempting to make a rule for developers to abide by.
This is effectively saying "when alt is needed and not available don't
put null alt just to pass the validator". This is in fact pretty
interesting. AT already applies heuristics to guess at the appropriate
representation for the image, but not _image_ heuristics. This would be
nice, but we're not there yet. This could be a nice feature for
repair tools since they could easily identify images that are already
determined by the developer as needing an alt.

At the end of the day, it isn't going to make a big difference either
way. Images need useful alt values in a well-defined set of
circumstances. Whether the image alt attribute is missing or it is
alt="" for an image that is important, there is an accessibility
problem, and one that would need to be verified even if a page passes a
validator.

In the page Steve posted, the images are links. In the HTML5 document,
this case is covered in an example, but not specifically mentioned. In
their image page example, the images don't seem to be links. I'll
certainly agree that this requires clarification, but if the images that
were links were missing alt this is still and accessibility issue and
some amount of manual verification is still required to be assured of
compliance.

AWK


From: John Foliot - Stanford Online Accessibility Program
Date: Thu, Aug 30 2007 8:30PM
Subject: Re: Investigating the proposedaltattributerecommendationsin HTML 5
← Previous message | No next message

Andrew Kirkpatrick wrote:
>
> To me this is attempting to make a rule for developers to abide by.
> This is effectively saying "when alt is needed and not available
> don't put null alt just to pass the validator".

Instead, "...the alt attribute may be omitted..." - Andrew, how can this be
of any benefit to accessibility concerns?

> This is in fact
> pretty interesting. AT already applies heuristics to guess at the
> appropriate representation for the image, but not _image_ heuristics.
> This would be
> nice, but we're not there yet.

Precisely, thus it is the responsibility of the developer to bridge the gap,
and one simple way is to provide the alternative text. Otherwise,
currently, the heuristic result often ends up being "34598_6352.jpg".
Providing the option of omitting alt text will be used as a "Free Pass" by
some, and they will point to the spec and say that it's allowed. It's the
thin edge of the wedge.

> This could be a nice feature for
> repair tools since they could easily identify images that are already
> determined by the developer as needing an alt.

As does a search for alt=""...

>
> At the end of the day, it isn't going to make a big difference either
> way. Images need useful alt values in a well-defined set of
> circumstances. Whether the image alt attribute is missing or it is
> alt="" for an image that is important, there is an accessibility
> problem, and one that would need to be verified even if a page passes
> a validator.

Here we are in total agreement. However, entrenching the concept of
"optional" alternative text will be abused and mis-used in as many or more
instances than the very few examples currently being suggested (and BTW, who
uploads 3,000 photos of their vacation?) Again, once the door has been
wedged open, it will be difficult to close Pandora's Box.

JF