WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: Creating Valid Code

for

Number of posts in this thread: 10 (In chronological order)

From: Ryan E. Benson
Date: Thu, Sep 08 2011 6:51PM
Subject: Creating Valid Code
No previous message | Next message →

I was asked to look at part of a site. Most of the pages had over 20
validation errors. While the developer hasn't been creating sites for
too long, my lead told me to essentially ignore them. He says
non-valid code is kind of the standard these days "in the real world."
I know a good amount of people on the list either work for large
organizations/companies, is this just how it is or should I be pushing
for valid code. I would be fine with the errors if they were under
five or so, but 20 is little high for my blood.

What do you guys think?

--
Ryan E. Benson

From: Birkir R. Gunnarsson
Date: Thu, Sep 08 2011 7:03PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

Ryan

Can you post a sample of these errors, or tell us what the website is}
I agree that 20 is a bit high. For a test that Anec, the Euroean
consumer organiation carried out using WCAG 2.0 AA compliance, they
allowd up to 10 erros before they started taking points off, seemed
somewhat arbitrary, but I know the lady in charge of the study and she
is very smart nd thorrough.
I think it does depent, at lesat somewhat, on the types of errors you
are getting. I am fairly new to this business myself, and I try to
focus on themes in the types of errors I am finding, and emphasize
that to the web developers. If the problem is Javascript, I send tem
links on examples of accessible and inaccessible Javascript, along
with some comments about how it affects my experience as a screen
reader user. If it is failure to use alt text appropriately, I do the
same for that.
If someone has better approaches, I am fairly new and open to ideas
and learning more myself.
Hope these thoughts provide some initial, well, fod for tought. <(
=B

On 9/9/11, Ryan E. Benson < = EMAIL ADDRESS REMOVED = > wrote:
> I was asked to look at part of a site. Most of the pages had over 20
> validation errors. While the developer hasn't been creating sites for
> too long, my lead told me to essentially ignore them. He says
> non-valid code is kind of the standard these days "in the real world."
> I know a good amount of people on the list either work for large
> organizations/companies, is this just how it is or should I be pushing
> for valid code. I would be fine with the errors if they were under
> five or so, but 20 is little high for my blood.
>
> What do you guys think?
>
> --
> Ryan E. Benson
>

From: John Foliot
Date: Thu, Sep 08 2011 7:33PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

Ryan E. Benson wrote:
>
> I was asked to look at part of a site. Most of the pages had over 20
> validation errors. While the developer hasn't been creating sites for
> too long, my lead told me to essentially ignore them. He says
> non-valid code is kind of the standard these days "in the real world."
> I know a good amount of people on the list either work for large
> organizations/companies, is this just how it is or should I be pushing
> for valid code. I would be fine with the errors if they were under
> five or so, but 20 is little high for my blood.
>
> What do you guys think?

Hi Ryan,

Code validation, in-and-of-itself, may have very little impact on true
accessibility: I've seen web pages that passed all the technical
validation requirements and still turn out a bit of an accessibility mess,
and conversely I've seen pages that do not pass validation be highly
useable and accessible. It can also depend on which DTD (or lack of in
HTML5) you are validating against: for example any document in HTML4 or
XHTML1 that contains ARIA will not pass a mechanical validator, yet
removing ARIA to meet validation requirements seems something of a
backward step, don't you think?

As Birkir has suggested, the kinds of validation errors being presented
are generally more important than a sum-total of errors, and like most
other aspects of web accessibility it takes some reasoning to understand
the impact of validation versus non-validation. By practice, I use
validation reports today as something of my Canary in the Coal Mine - if
there are a large number of validation errors then I am fairly confident
that there will be accessibility issues as well, as it is usually
symptomatic of a big problem. However if all the errors are being
generated by, for example, un-escaped ampersands (&), then the real impact
on accessibility is pretty much negligible today (thanks to browser error
recovery).

So I think that the answer you seek is somewhere in the middle of what you
are thinking and what your boss might be suggesting. Best suggestion is
to examine the errors and see what type of errors are being generated, and
then make your evaluation based upon that. Don't get caught in a numbers
game, stay focused on real results and user-impacts. I challenge
developers to strive for validation (as achieving it means being focused
on attention to detail, which generally also means better accessibility -
but not always), but when it comes to evaluating 'success', validation has
a lower value in the greater picture: nice to have but not critical.

HTH

JF
============================
John  Foliot
Program Manager
Stanford Online Accessibility Program
http://soap.stanford.edu
Stanford University
Tel: 650-468-5785

---
Co-chair - W3C HTML5 Accessibility Task Force (Media)
http://www.w3.org/WAI/PF/HTML/wiki/Main_Page

============================

From: Ryan E. Benson
Date: Thu, Sep 08 2011 7:39PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

Birkir,

I cannot post a link to the site since it is internal. I checked the
site using the W3C validator, and a lot were ending the tags in >
versus />. The pages don't use javascript at all really, in fact one
of the blocks of Javascript are not being used anymore. As I said the
developer is kind of new and asked me if a <noscript> was needed. It
is a script to pass the URL to the contact page, but the contact page
doesn't pass the URL to a field on the page or e-mail it sends. So a
<noscript> is needed.

The pages are not complicated, usually a h1, then a few paragraphs,
and a nav bar.

--
Ryan E. Benson



On Thu, Sep 8, 2011 at 9:01 PM, Birkir R. Gunnarsson
< = EMAIL ADDRESS REMOVED = > wrote:
> Ryan
>
> Can you post a sample of these errors, or tell us what the website is}
> I agree that 20 is a bit high. For a test that Anec, the Euroean
> consumer organiation carried out using WCAG 2.0 AA compliance, they
> allowd up to 10 erros before they started taking points off, seemed
> somewhat arbitrary, but I know the lady in charge of the study and she
> is very smart nd thorrough.
> I think it does depent, at lesat somewhat, on the types of errors you
> are getting. I am fairly new to this business myself, and I try to
> focus on themes in the types of errors I am finding, and emphasize
> that to the web developers. If the problem is Javascript, I send tem
> links on examples of accessible and inaccessible Javascript, along
> with some comments about how it affects my experience as a screen
> reader user. If it is failure to use alt text appropriately, I do the
> same for that.
> If someone has better approaches, I am fairly new and open to ideas
> and learning more myself.
> Hope these thoughts provide some initial, well, fod for tought. <(
> =B
>
> On 9/9/11, Ryan E. Benson < = EMAIL ADDRESS REMOVED = > wrote:
>> I was asked to look at part of a site. Most of the pages had over 20
>> validation errors. While the developer hasn't been creating sites for
>> too long, my lead told me to essentially ignore them. He says
>> non-valid code is kind of the standard these days "in the real world."
>> I know a good amount of people on the list either work for large
>> organizations/companies, is this just how it is or should I be pushing
>> for valid code.  I would be fine with the errors if they were under
>> five or so, but 20 is little high for my blood.
>>
>> What do you guys think?
>>
>> --
>> Ryan E. Benson
>>

From: Ryan E. Benson
Date: Thu, Sep 08 2011 8:18PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

Hi John,

> for example any document in HTML4 or
> XHTML1 that contains ARIA will not pass a mechanical validator, yet
> removing ARIA to meet validation requirements seems something of a
> backward step, don't you think?
As an organization, we are just now talking about ARIA. In fact I was
asked by a mid-level web guy asked me if this ARIA thing is any good.
So I have a few months to worry about that. Anyhow, I get your point.

I agree with you that validity doesn't quite measure accessibility. I
guess I am seeking if there is a time to throw the code and hit
restart. Another option is to re-educate.

--
Ryan E. Benson



On Thu, Sep 8, 2011 at 9:33 PM, John Foliot < = EMAIL ADDRESS REMOVED = > wrote:
> Ryan E. Benson wrote:
>>
>> I was asked to look at part of a site. Most of the pages had over 20
>> validation errors. While the developer hasn't been creating sites for
>> too long, my lead told me to essentially ignore them. He says
>> non-valid code is kind of the standard these days "in the real world."
>> I know a good amount of people on the list either work for large
>> organizations/companies, is this just how it is or should I be pushing
>> for valid code.  I would be fine with the errors if they were under
>> five or so, but 20 is little high for my blood.
>>
>> What do you guys think?
>
> Hi Ryan,
>
> Code validation, in-and-of-itself, may have very little impact on true
> accessibility: I've seen web pages that passed all the technical
> validation requirements and still turn out a bit of an accessibility mess,
> and conversely I've seen pages that do not pass validation be highly
> useable and accessible. It can also depend on which DTD (or lack of in
> HTML5) you are validating against: for example any document in HTML4 or
> XHTML1 that contains ARIA will not pass a mechanical validator, yet
> removing ARIA to meet validation requirements seems something of a
> backward step, don't you think?
>
> As Birkir has suggested, the kinds of validation errors being presented
> are generally more important than a sum-total of errors, and like most
> other aspects of web accessibility it takes some reasoning to understand
> the impact of validation versus non-validation. By practice, I use
> validation reports today as something of my Canary in the Coal Mine - if
> there are a large number of validation errors then I am fairly confident
> that there will be accessibility issues as well, as it is usually
> symptomatic of a big problem. However if all the errors are being
> generated by, for example, un-escaped ampersands (&), then the real impact
> on accessibility is pretty much negligible today (thanks to browser error
> recovery).
>
> So I think that the answer you seek is somewhere in the middle of what you
> are thinking and what your boss might be suggesting.  Best suggestion is
> to examine the errors and see what type of errors are being generated, and
> then make your evaluation based upon that. Don't get caught in a numbers
> game, stay focused on real results and user-impacts. I challenge
> developers to strive for validation (as achieving it means being focused
> on attention to detail, which generally also means better accessibility -
> but not always), but when it comes to evaluating 'success', validation has
> a lower value in the greater picture: nice to have but not critical.
>
> HTH
>
> JF
> ============================
> John  Foliot
> Program Manager
> Stanford Online Accessibility Program
> http://soap.stanford.edu
> Stanford University
> Tel: 650-468-5785
>
> ---
> Co-chair - W3C HTML5 Accessibility Task Force (Media)
> http://www.w3.org/WAI/PF/HTML/wiki/Main_Page
>
> ============================
>
>
>
>

From: Elle
Date: Thu, Sep 08 2011 9:24PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

I meant to add this small footnote:

* we also used that opportunity to build in some usability requirements to
get closer to the true goal for accessibility :)



On Thu, Sep 8, 2011 at 11:24 PM, Elle < = EMAIL ADDRESS REMOVED = > wrote:

> Ryan:
>
> If it helps, I can give you my perspective at a large company with over
> 30,000 employees. Until just last year, we were struggling to justify the
> effort to insist on W3C valid code. Most of the concerns involved increased
> development costs, additional testing, and blowing up project time lines.
> Keeping in mind that not all companies have the same challenges, most large
> organizations also outsource a great deal of development to contracted labor
> who may or may not be familiar with W3C standards. So, there's an undeniable
> cost to establishing web standards like W3C valid markup. And what did it
> really get us? That was the main question that needed answering.
>
> When we were building our web accessibility program last year, we decided
> to use that moment to incorporate W3C validity* into our accessibility
> requirements. As John said, "code validation, in-and-of-itself, may have
> very little impact on true accessibility." But, it does represent a
> paradigm shift for IT teams: semantic markup matters. Previously, like many
> fast moving big companies, we would build (largely ASP.NET) web
> applications rapidly with wireframes and visual mock-ups for agile
> requirements. There wasn't much consideration or scrutiny over what was
> "under the hood" when it came time to show our business partners what we had
> created for them, as long as it performed as was requested. Web
> accessibility changed all that, as it has a way of getting up close and
> personal with an individual's source code. Accessibility requires
> transparency. So, while we're under the hood, why not create a stronger
> foundation for cross-browser operability and device independence? We figured
> if we did it right, we wouldn't have to continually chase the latest browser
> versions with code updates.
>
> Since that time, we've found that creating this base level requirement (all
> code must validate through W3C with a testing artifact that's loaded into
> our Software Development Life Cycle) has streamlined our development process
> and reduced the pesky defects we used to encounter during user acceptance
> testing. That's a quantifiable cost savings. Progressive enhancement and
> MVC as a design pattern have made that easier to achieve and even cheaper.
> We do have exceptions for third party source code (example: vendors who
> supply scripts), but that's a different battle to fight.
>
> Hope that helps,
> Elle
>
> --
> If you want to build a ship, don't drum up the people to gather wood,
> divide the work, and give orders. Instead, teach them to yearn for the vast
> and endless sea.
> - Antoine De Saint-Exupéry, The Little Prince
>
>


--
If you want to build a ship, don't drum up the people to gather wood, divide
the work, and give orders. Instead, teach them to yearn for the vast and
endless sea.
- Antoine De Saint-Exupéry, The Little Prince

From: Elle
Date: Thu, Sep 08 2011 9:30PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

Ryan:

If it helps, I can give you my perspective at a large company with over
30,000 employees. Until just last year, we were struggling to justify the
effort to insist on W3C valid code. Most of the concerns involved increased
development costs, additional testing, and blowing up project time lines.
Keeping in mind that not all companies have the same challenges, most large
organizations also outsource a great deal of development to contracted labor
who may or may not be familiar with W3C standards. So, there's an undeniable
cost to establishing web standards like W3C valid markup. And what did it
really get us? That was the main question that needed answering.

When we were building our web accessibility program last year, we decided to
use that moment to incorporate W3C validity* into our accessibility
requirements. As John said, "code validation, in-and-of-itself, may have
very little impact on true accessibility." But, it does represent a paradigm
shift for IT teams: semantic markup matters. Previously, like many fast
moving big companies, we would build (largely ASP.NET) web applications
rapidly with wireframes and visual mock-ups for agile requirements. There
wasn't much consideration or scrutiny over what was "under the hood" when it
came time to show our business partners what we had created for them, as
long as it performed as was requested. Web accessibility changed all that,
as it has a way of getting up close and personal with an individual's source
code. Accessibility requires transparency. So, while we're under the hood,
why not create a stronger foundation for cross-browser operability and
device independence? We figured if we did it right, we wouldn't have to
continually chase the latest browser versions with code updates.

Since that time, we've found that creating this base level requirement (all
code must validate through W3C with a testing artifact that's loaded into
our Software Development Life Cycle) has streamlined our development process
and reduced the pesky defects we used to encounter during user acceptance
testing. That's a quantifiable cost savings. Progressive enhancement and
MVC as a design pattern have made that easier to achieve and even cheaper.
We do have exceptions for third party source code (example: vendors who
supply scripts), but that's a different battle to fight.

Hope that helps,
Elle

--
If you want to build a ship, don't drum up the people to gather wood, divide
the work, and give orders. Instead, teach them to yearn for the vast and
endless sea.
- Antoine De Saint-Exupéry, The Little Prince

From: Lucy Greco
Date: Fri, Sep 09 2011 9:54AM
Subject: Re: Creating Valid Code
← Previous message | Next message →

I would say make sure the pages work for a user and if they do maybe the hits are ok then

Lucy Greco
Assistive Technology Specialist
Disabled Student's Program UC Berkeley
(510) 643-7591
http://attlc.berkeley.edu
http://webaccess.berkeley.edu


-----Original Message-----
From: = EMAIL ADDRESS REMOVED = [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Ryan E. Benson
Sent: Thursday, September 08, 2011 5:53 PM
To: WebAIM Discussion List
Subject: [WebAIM] Creating Valid Code

I was asked to look at part of a site. Most of the pages had over 20 validation errors. While the developer hasn't been creating sites for too long, my lead told me to essentially ignore them. He says non-valid code is kind of the standard these days "in the real world."
I know a good amount of people on the list either work for large organizations/companies, is this just how it is or should I be pushing for valid code. I would be fine with the errors if they were under five or so, but 20 is little high for my blood.

What do you guys think?

--
Ryan E. Benson

From: Barry Johnson
Date: Fri, Sep 09 2011 12:06PM
Subject: Re: Creating Valid Code
← Previous message | Next message →

I have to agree strongly with Elle.

While valid code does not = web site accessibility, creating W3C valid code
helps make your site available to all browsers on all platforms, and this
means a lot less work has browsers and platforms change in the future. It
also allows the code to read more easily by the assitive technology, as they
only have the code standards to create their tools.
The fact that this get the developers up close and personal with the code
and helps them make the right decisions.

Code to the standards and your code will be available to any
technology/platform.

Barry


On Thu, Sep 8, 2011 at 11:25 PM, Elle < = EMAIL ADDRESS REMOVED = > wrote:

> I meant to add this small footnote:
>
> * we also used that opportunity to build in some usability requirements to
> get closer to the true goal for accessibility :)
>
>
>
> On Thu, Sep 8, 2011 at 11:24 PM, Elle < = EMAIL ADDRESS REMOVED = > wrote:
>
> > Ryan:
> >
> > If it helps, I can give you my perspective at a large company with over
> > 30,000 employees. Until just last year, we were struggling to justify
> the
> > effort to insist on W3C valid code. Most of the concerns involved
> increased
> > development costs, additional testing, and blowing up project time lines.
> > Keeping in mind that not all companies have the same challenges, most
> large
> > organizations also outsource a great deal of development to contracted
> labor
> > who may or may not be familiar with W3C standards. So, there's an
> undeniable
> > cost to establishing web standards like W3C valid markup. And what did it
> > really get us? That was the main question that needed answering.
> >
> > When we were building our web accessibility program last year, we decided
> > to use that moment to incorporate W3C validity* into our accessibility
> > requirements. As John said, "code validation, in-and-of-itself, may have
> > very little impact on true accessibility." But, it does represent a
> > paradigm shift for IT teams: semantic markup matters. Previously, like
> many
> > fast moving big companies, we would build (largely ASP.NET) web
> > applications rapidly with wireframes and visual mock-ups for agile
> > requirements. There wasn't much consideration or scrutiny over what was
> > "under the hood" when it came time to show our business partners what we
> had
> > created for them, as long as it performed as was requested. Web
> > accessibility changed all that, as it has a way of getting up close and
> > personal with an individual's source code. Accessibility requires
> > transparency. So, while we're under the hood, why not create a stronger
> > foundation for cross-browser operability and device independence? We
> figured
> > if we did it right, we wouldn't have to continually chase the latest
> browser
> > versions with code updates.
> >
> > Since that time, we've found that creating this base level requirement
> (all
> > code must validate through W3C with a testing artifact that's loaded into
> > our Software Development Life Cycle) has streamlined our development
> process
> > and reduced the pesky defects we used to encounter during user acceptance
> > testing. That's a quantifiable cost savings. Progressive enhancement and
> > MVC as a design pattern have made that easier to achieve and even
> cheaper.
> > We do have exceptions for third party source code (example: vendors who
> > supply scripts), but that's a different battle to fight.
> >
> > Hope that helps,
> > Elle
> >
> > --
> > If you want to build a ship, don't drum up the people to gather wood,
> > divide the work, and give orders. Instead, teach them to yearn for the
> vast
> > and endless sea.
> > - Antoine De Saint-Exupéry, The Little Prince
> >
> >
>
>
> --
> If you want to build a ship, don't drum up the people to gather wood,
> divide
> the work, and give orders. Instead, teach them to yearn for the vast and
> endless sea.
> - Antoine De Saint-Exupéry, The Little Prince
>

From: Bourne, Sarah (ITD)
Date: Wed, Sep 14 2011 9:54AM
Subject: Re: Creating Valid Code
← Previous message | No next message

Ryan,

Elaborating on what Elle and John said, you can get quirky problems if you don't have valid code. This is because the browsers have to guess what you meant. This can result in things like JAWS reading a form correctly one time but not the next time you check. If you are getting inconsistent results with AT, you should go back and look at those HTML errors the validator found. Tags that aren't closed and/or that are nested improperly are likely suspects.

sb

Sarah E. Bourne
Director of Assistive Technology &
Mass.Gov Chief Technology Strategist
Information Technology Division
Commonwealth of Massachusetts
1 Ashburton Pl. rm 1601 Boston MA 02108
617-626-4502  fax 617-626-4516
http://twitter.com/sarahebourne
= EMAIL ADDRESS REMOVED =
http://www.mass.gov/itd


-----Original Message-----
From: = EMAIL ADDRESS REMOVED = [mailto: = EMAIL ADDRESS REMOVED = ] On Behalf Of Ryan E. Benson
Sent: Thursday, September 08, 2011 8:53 PM
To: WebAIM Discussion List
Subject: [WebAIM] Creating Valid Code

I was asked to look at part of a site. Most of the pages had over 20 validation errors. While the developer hasn't been creating sites for too long, my lead told me to essentially ignore them. He says non-valid code is kind of the standard these days "in the real world."
I know a good amount of people on the list either work for large organizations/companies, is this just how it is or should I be pushing for valid code. I would be fine with the errors if they were under five or so, but 20 is little high for my blood.

What do you guys think?

--
Ryan E. Benson