WebAIM - Web Accessibility In Mind

E-mail List Archives

Thread: Does 1.4.10 essentially replace 1.4.4?

for

Number of posts in this thread: 9 (In chronological order)

From: glen walker
Date: Sat, Sep 04 2021 8:26AM
Subject: Does 1.4.10 essentially replace 1.4.4?
No previous message | Next message →

1.4.4 requires zooming to 200% "without loss of content or functionality".
1.4.10 requires zooming to 400% "without loss of information or
functionality"

I'm not sure why the wording is different between "loss of *content*" vs
"loss of *information*". I suppose the latter could lose DOM elements
(which I consider content) but the general gist of information is still
there so it might pass. But that's a side topic. For all intents and
purposes, they're both talking about not losing stuff when zoomed.

1.4.4 doesn't care what size your browser window is so in theory you could
zoom to 200% and lose content if your browser is small but if you maximize
the browser and you have a large monitor, you might not lose any content so
it could pass.

1.4.10 says to test with a 1280 width browser (for vertical content).

So when testing, do you really need to test for 1.4.4 if you're trying to
pass WCAG 2.1 AA conformance? Can't you just test for 1.4.10 and if that
passes, you also pass 1.4.4?

I suppose, in theory again, you might have different CSS breakpoints and
the breakpoint for 400% zoom might look good but the breakpoint for 200%
might not. I've never had that happen but it is possible so testing at
both 200% and 400% is probably wise.

I'm just tossing a discussion topic out there. If you're in a hurry to do
a quick scan, you might be able to test for 400% and check off both success
criteria. Of course, when pressing Ctrl++, you have to go through 200%
before 400% (unless you've modified your browser zoom to jump right to
400%) so it's only a few seconds extra to pause at 200%, scroll the page to
see if it looks good, then continue up to 400%.

I've also glossed over the scrollbar requirement for 1.4.10.

What I typically do is zoom to 200%, see if it looks good at my current
browser size, then continue up to 400% and resize my browser to 1280 (using
responsive view from the code inspector [both firefox and chrome], which
makes it easy to set the browser width) and see if it looks good and no
horizontal scrollbar appears.

From: Birkir R. Gunnarsson
Date: Sat, Sep 04 2021 8:42AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

This is my understanding too.
I find this article from Knowbility helpful in trying to make sense of 1.4.10
https://knowbility.org/blog/2018/WCAG21-1410Reflow


On 9/4/21, glen walker < = EMAIL ADDRESS REMOVED = > wrote:
> 1.4.4 requires zooming to 200% "without loss of content or functionality".
> 1.4.10 requires zooming to 400% "without loss of information or
> functionality"
>
> I'm not sure why the wording is different between "loss of *content*" vs
> "loss of *information*". I suppose the latter could lose DOM elements
> (which I consider content) but the general gist of information is still
> there so it might pass. But that's a side topic. For all intents and
> purposes, they're both talking about not losing stuff when zoomed.
>
> 1.4.4 doesn't care what size your browser window is so in theory you could
> zoom to 200% and lose content if your browser is small but if you maximize
> the browser and you have a large monitor, you might not lose any content so
> it could pass.
>
> 1.4.10 says to test with a 1280 width browser (for vertical content).
>
> So when testing, do you really need to test for 1.4.4 if you're trying to
> pass WCAG 2.1 AA conformance? Can't you just test for 1.4.10 and if that
> passes, you also pass 1.4.4?
>
> I suppose, in theory again, you might have different CSS breakpoints and
> the breakpoint for 400% zoom might look good but the breakpoint for 200%
> might not. I've never had that happen but it is possible so testing at
> both 200% and 400% is probably wise.
>
> I'm just tossing a discussion topic out there. If you're in a hurry to do
> a quick scan, you might be able to test for 400% and check off both success
> criteria. Of course, when pressing Ctrl++, you have to go through 200%
> before 400% (unless you've modified your browser zoom to jump right to
> 400%) so it's only a few seconds extra to pause at 200%, scroll the page to
> see if it looks good, then continue up to 400%.
>
> I've also glossed over the scrollbar requirement for 1.4.10.
>
> What I typically do is zoom to 200%, see if it looks good at my current
> browser size, then continue up to 400% and resize my browser to 1280 (using
> responsive view from the code inspector [both firefox and chrome], which
> makes it easy to set the browser width) and see if it looks good and no
> horizontal scrollbar appears.
> > > > >


--
Work hard. Have fun. Make history.

From: Patrick H. Lauke
Date: Sat, Sep 04 2021 8:45AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

On 04/09/2021 15:26, glen walker wrote:
> 1.4.4 requires zooming to 200% "without loss of content or functionality".
> 1.4.10 requires zooming to 400% "without loss of information or
> functionality"

Quickfire answer: if you zoom to 200% (at whatever screen size - and the
fact that this is not clearly defined anywhere is a separate problem)
and it causes horizontal and vertical scrollbars, that's OK for 1.4.4 as
long as content is still reachable/not cut off/overlapped.

But if you zoom to exactly 320 CSS px width / 256 CSS px height, and you
get either content that's cut off etc, OR you get horizontal and
vertical scrollbars (that mean you have to scroll back and forth to read
a chunk of content, in both directions), then that FAILS 1.4.10

see 1.4.10 as a stricter version of 1.4.4 (due to the scrollbars being a
no-no in 1.4.10)


--
Patrick H. Lauke

https://www.splintered.co.uk/ | https://github.com/patrickhlauke
https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: glen walker
Date: Sun, Sep 05 2021 11:22AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

Thanks Birkir and Patrick. I understand the fundamentals of both success
criteria. My question was more academic in nature. If you want to conform
to WCAG 2.1 AA then you must pass 1.4.10. If you pass 1.4.10, doesn't that
(mostly) imply you also pass 1.4.4 so there's no need to test for 1.4.4?

If you look good at 400% (and you only have one scrollbar) then you'll
probably look good at 200% (regardless of scrollbars).

We can create examples where this isn't true, of course, such as Patrick's
example in https://patrickhlauke.github.io/wcag-interpretation/?full#58

> if you zoom to exactly 320 CSS px width / 256 CSS px height

That's another aspect that isn't completely clear (to me). 1.4.10 doesn't
really say that *both* the width and height must be set. It just says that
vertical scrolling content (such as English, German, Spanish, etc) doesn't
require scrolling in two dimensions when at (exactly) 320 CSS px width. It
doesn't say I must set the height too.

Same with horizontal scrolling content (such as Mandarin, Korean, Arabic).
The requirement refers to setting the height to 256 CSS px but doesn't say
I must set the width too.

Conceptually, if you think of the vertical scrolling case, if you don't
have a horizontal scrollbar then you're good for 1.4.10 and since you're
scrolling vertically, it doesn't really matter what the height of the
browser is.

From: Patrick H. Lauke
Date: Sun, Sep 05 2021 11:30AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

On 05/09/2021 18:22, glen walker wrote:
> Thanks Birkir and Patrick. I understand the fundamentals of both success
> criteria. My question was more academic in nature. If you want to conform
> to WCAG 2.1 AA then you must pass 1.4.10. If you pass 1.4.10, doesn't that
> (mostly) imply you also pass 1.4.4 so there's no need to test for 1.4.4?

They're independent, but yes if you start off testing 1.4.10 and it's
fine, it's likely that 1.4.4 passes as well (barring oddity of
intermediate viewport sizes if the developers have only set specific
ones and zooming to 200% would fall between the gaps and leave you in
unforeseen broken dimensions.

> That's another aspect that isn't completely clear (to me). 1.4.10 doesn't
> really say that *both* the width and height must be set. It just says that
> vertical scrolling content (such as English, German, Spanish, etc) doesn't
> require scrolling in two dimensions when at (exactly) 320 CSS px width. It
> doesn't say I must set the height too.
>
> Same with horizontal scrolling content (such as Mandarin, Korean, Arabic).
> The requirement refers to setting the height to 256 CSS px but doesn't say
> I must set the width too.
>
> Conceptually, if you think of the vertical scrolling case, if you don't
> have a horizontal scrollbar then you're good for 1.4.10 and since you're
> scrolling vertically, it doesn't really matter what the height of the
> browser is.

As the idea behind 1.4.10 was a desktop user on a 1280x768 display
zooming to 400%, it makes sense to test at that exact dimension. And
regardless of language, there might be situations (e.g. with fixed
headers/footer/etc) where there's no bi-directional scrollbars, but
where content is completely covered by other fixed/absolutely positioned
content. Or situations like a modal dialog that is centered and gets cut
off at the top and bottom, with no way for the user to scroll that modal
properly into view. Those are more general "without loss of information
or functionality" failures. (But yes this aspect is still quite vague in
1.4.10 and the understanding doc, and I'm sure some will argue that
those aren't failures of 1.4.10 at all)

P
--
Patrick H. Lauke

https://www.splintered.co.uk/ | https://github.com/patrickhlauke
https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Steve Green
Date: Sun, Sep 05 2021 2:38PM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

I have seen plenty of websites where the media queries are messed up such that 1.4.10 passes, but 1.4.4 doesn't. The desktop and mobile layouts (corresponding to 100% and 400% zoom) have generally been done properly, but the intermediate breakpoints for tablets are often buggy.

It's also worth noting that zooming is only one of the ways that a website can pass 1.4.4. It is possible (albeit unlikely) that a website relies on a text resizing widget to meet 1.4.4 because zooming doesn't work properly at 200% for some reason.

These sort of issues are less likely to happen with new websites, but are perhaps more likely with older fixed-width websites that have had some media queries added to make them responsive. We recently worked on a 15 year-old web application created in XHTML 1.0 where the client was trying to do this.

Steve Green
Managing Director
Test Partners Ltd


-----Original Message-----
From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > On Behalf Of Patrick H. Lauke
Sent: 05 September 2021 18:31
To: = EMAIL ADDRESS REMOVED =
Subject: Re: [WebAIM] Does 1.4.10 essentially replace 1.4.4?

On 05/09/2021 18:22, glen walker wrote:
> Thanks Birkir and Patrick. I understand the fundamentals of both
> success criteria. My question was more academic in nature. If you
> want to conform to WCAG 2.1 AA then you must pass 1.4.10. If you pass
> 1.4.10, doesn't that
> (mostly) imply you also pass 1.4.4 so there's no need to test for 1.4.4?

They're independent, but yes if you start off testing 1.4.10 and it's fine, it's likely that 1.4.4 passes as well (barring oddity of intermediate viewport sizes if the developers have only set specific ones and zooming to 200% would fall between the gaps and leave you in unforeseen broken dimensions.

> That's another aspect that isn't completely clear (to me). 1.4.10
> doesn't really say that *both* the width and height must be set. It
> just says that vertical scrolling content (such as English, German,
> Spanish, etc) doesn't require scrolling in two dimensions when at
> (exactly) 320 CSS px width. It doesn't say I must set the height too.
>
> Same with horizontal scrolling content (such as Mandarin, Korean, Arabic).
> The requirement refers to setting the height to 256 CSS px but doesn't
> say I must set the width too.
>
> Conceptually, if you think of the vertical scrolling case, if you
> don't have a horizontal scrollbar then you're good for 1.4.10 and
> since you're scrolling vertically, it doesn't really matter what the
> height of the browser is.

As the idea behind 1.4.10 was a desktop user on a 1280x768 display zooming to 400%, it makes sense to test at that exact dimension. And regardless of language, there might be situations (e.g. with fixed
headers/footer/etc) where there's no bi-directional scrollbars, but where content is completely covered by other fixed/absolutely positioned content. Or situations like a modal dialog that is centered and gets cut off at the top and bottom, with no way for the user to scroll that modal properly into view. Those are more general "without loss of information or functionality" failures. (But yes this aspect is still quite vague in
1.4.10 and the understanding doc, and I'm sure some will argue that those aren't failures of 1.4.10 at all)

P
--
Patrick H. Lauke

https://www.splintered.co.uk/ | https://github.com/patrickhlauke https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
twitter: @patrick_h_lauke | skype: patrick_h_lauke

From: Mallory
Date: Mon, Sep 06 2021 2:08AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

...there's also the cute wording between the two.

"up to 200%"
"at 400%"

With those soulless clients who are lawyer-driven to get a "WCAG audit", they can pass 200%, 400% and fail huge and hard at 300% and "pass" WCAG. Eric Eggert and Pat and maybe a few others have pointed out that if 1.4.10 was "up to" then with all the gazillion possible viewport sizes and breakpoints, it could in theory become impossible to test.

Also the 200% one I've read as saying that regardless of what size you start at by default, the user can make the text twice as big. The 400% one doesn't, it says content reflows properly. I have run into sites being quite evil, where they found art direction more important than usability, where if you used browser zoom to increase things the actual font glyphs remained the same size (they would use viewportheight/viewportwidth units for the fonts, ug). Such a site could in theory pass the 400% (if it reflows properly) but fail 200% (I cannot make the text twice as big).

When auditing, if I'm not pressed for time, I'll check 200% in Firefox using its text-only zoom and test 400% in a blink with the browser sized to 1280 width, and then for sh*ts check portrait on my Android potato. Most fails I run into are due to devs making phone-portrait reflow nicely but due to stickies makes 400% impossible on a laptop since laptops are basically landscape orientation.
(No, I can't fail a site for breaking with text-only zoom but I have discovered it's so closely related to possible problems when checking the Text Spacing SC that it's good to just see what you get with it; I check Text Spacing coupled with increased zoom as well. They *should* work well together.)

cheers,
_mallory

On Sun, Sep 5, 2021, at 10:38 PM, Steve Green wrote:
> I have seen plenty of websites where the media queries are messed up
> such that 1.4.10 passes, but 1.4.4 doesn't. The desktop and mobile
> layouts (corresponding to 100% and 400% zoom) have generally been done
> properly, but the intermediate breakpoints for tablets are often buggy.
>
> It's also worth noting that zooming is only one of the ways that a
> website can pass 1.4.4. It is possible (albeit unlikely) that a website
> relies on a text resizing widget to meet 1.4.4 because zooming doesn't
> work properly at 200% for some reason.
>
> These sort of issues are less likely to happen with new websites, but
> are perhaps more likely with older fixed-width websites that have had
> some media queries added to make them responsive. We recently worked on
> a 15 year-old web application created in XHTML 1.0 where the client was
> trying to do this.
>
> Steve Green
> Managing Director
> Test Partners Ltd
>
>
> -----Original Message-----
> From: WebAIM-Forum < = EMAIL ADDRESS REMOVED = > On Behalf Of
> Patrick H. Lauke
> Sent: 05 September 2021 18:31
> To: = EMAIL ADDRESS REMOVED =
> Subject: Re: [WebAIM] Does 1.4.10 essentially replace 1.4.4?
>
> On 05/09/2021 18:22, glen walker wrote:
> > Thanks Birkir and Patrick. I understand the fundamentals of both
> > success criteria. My question was more academic in nature. If you
> > want to conform to WCAG 2.1 AA then you must pass 1.4.10. If you pass
> > 1.4.10, doesn't that
> > (mostly) imply you also pass 1.4.4 so there's no need to test for 1.4.4?
>
> They're independent, but yes if you start off testing 1.4.10 and it's
> fine, it's likely that 1.4.4 passes as well (barring oddity of
> intermediate viewport sizes if the developers have only set specific
> ones and zooming to 200% would fall between the gaps and leave you in
> unforeseen broken dimensions.
>
> > That's another aspect that isn't completely clear (to me). 1.4.10
> > doesn't really say that *both* the width and height must be set. It
> > just says that vertical scrolling content (such as English, German,
> > Spanish, etc) doesn't require scrolling in two dimensions when at
> > (exactly) 320 CSS px width. It doesn't say I must set the height too.
> >
> > Same with horizontal scrolling content (such as Mandarin, Korean, Arabic).
> > The requirement refers to setting the height to 256 CSS px but doesn't
> > say I must set the width too.
> >
> > Conceptually, if you think of the vertical scrolling case, if you
> > don't have a horizontal scrollbar then you're good for 1.4.10 and
> > since you're scrolling vertically, it doesn't really matter what the
> > height of the browser is.
>
> As the idea behind 1.4.10 was a desktop user on a 1280x768 display
> zooming to 400%, it makes sense to test at that exact dimension. And
> regardless of language, there might be situations (e.g. with fixed
> headers/footer/etc) where there's no bi-directional scrollbars, but
> where content is completely covered by other fixed/absolutely
> positioned content. Or situations like a modal dialog that is centered
> and gets cut off at the top and bottom, with no way for the user to
> scroll that modal properly into view. Those are more general "without
> loss of information or functionality" failures. (But yes this aspect is
> still quite vague in
> 1.4.10 and the understanding doc, and I'm sure some will argue that
> those aren't failures of 1.4.10 at all)
>
> P
> --
> Patrick H. Lauke
>
> https://www.splintered.co.uk/ | https://github.com/patrickhlauke
> https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
> > > archives at http://webaim.org/discussion/archives
> > > > > >

From: Mitchell Evan
Date: Tue, Sep 14 2021 2:09AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | Next message →

> As the idea behind 1.4.10 was a desktop user on a 1280x768 display
zooming to 400%, it makes sense to test at that exact dimension.

I agree we should test 1.4.10 in a common desktop display size at 400% (or
in a dev tools emulator equivalent). The most common aspect ratio of
desktop screens
<https://gs.statcounter.com/screen-resolution-stats/desktop/worldwide> is
1.78:1 with total usage over 60%, in screens such as 1280x720. Allowing for
browser and OS toolbars as shown in the Understanding 1.4.10 example video
<https://www.w3.org/WAI/WCAG21/Understanding/reflow.html#example-1-responsive-design>,
the
aspect ratio of the browser viewport (not the screen) is around 2:1.

>> 1.4.10 doesn't really say that *both* the width and height must be set.

True — it's a common misunderstanding
<https://github.com/act-rules/act-rules.github.io/issues/1318> that 1.4.10
Reflow specifies a 1280x1024 viewport — at most 3% usage, with a very
different aspect ratio of 1.25:1. The aspect ratio often makes a
pass-or-fail difference with real-world consequences, when sticky content
such as a fixed-position footer ends up covering the entire viewport.

Mitchell Evan, CPWA
linkedin.com/in/mitchellrevan <https://www.linkedin.com/in/mitchellrevan>
Twitter @mitchellrevan <https://twitter.com/mitchellrevan>
+49 1525 8950540
+1 510 375 6104

On Sun, 5 Sept 2021 at 20:00, < = EMAIL ADDRESS REMOVED = > wrote:

>
> ---------- Forwarded message ----------
> From: "Patrick H. Lauke" < = EMAIL ADDRESS REMOVED = >
> To: = EMAIL ADDRESS REMOVED =
> Cc:
> Bcc:
> Date: Sun, 5 Sep 2021 18:30:36 +0100
> Subject: Re: [WebAIM] Does 1.4.10 essentially replace 1.4.4?
> On 05/09/2021 18:22, glen walker wrote:
> > Thanks Birkir and Patrick. I understand the fundamentals of both success
> > criteria. My question was more academic in nature. If you want to
> conform
> > to WCAG 2.1 AA then you must pass 1.4.10. If you pass 1.4.10, doesn't
> that
> > (mostly) imply you also pass 1.4.4 so there's no need to test for 1.4.4?
>
> They're independent, but yes if you start off testing 1.4.10 and it's
> fine, it's likely that 1.4.4 passes as well (barring oddity of
> intermediate viewport sizes if the developers have only set specific
> ones and zooming to 200% would fall between the gaps and leave you in
> unforeseen broken dimensions.
>
> > That's another aspect that isn't completely clear (to me). 1.4.10
> doesn't
> > really say that *both* the width and height must be set. It just says
> that
> > vertical scrolling content (such as English, German, Spanish, etc)
> doesn't
> > require scrolling in two dimensions when at (exactly) 320 CSS px width.
> It
> > doesn't say I must set the height too.
> >
> > Same with horizontal scrolling content (such as Mandarin, Korean,
> Arabic).
> > The requirement refers to setting the height to 256 CSS px but doesn't
> say
> > I must set the width too.
> >
> > Conceptually, if you think of the vertical scrolling case, if you don't
> > have a horizontal scrollbar then you're good for 1.4.10 and since you're
> > scrolling vertically, it doesn't really matter what the height of the
> > browser is.
>
> As the idea behind 1.4.10 was a desktop user on a 1280x768 display
> zooming to 400%, it makes sense to test at that exact dimension. And
> regardless of language, there might be situations (e.g. with fixed
> headers/footer/etc) where there's no bi-directional scrollbars, but
> where content is completely covered by other fixed/absolutely positioned
> content. Or situations like a modal dialog that is centered and gets cut
> off at the top and bottom, with no way for the user to scroll that modal
> properly into view. Those are more general "without loss of information
> or functionality" failures. (But yes this aspect is still quite vague in
> 1.4.10 and the understanding doc, and I'm sure some will argue that
> those aren't failures of 1.4.10 at all)
>
> P
> --
> Patrick H. Lauke
>
> https://www.splintered.co.uk/ | https://github.com/patrickhlauke
> https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
> twitter: @patrick_h_lauke | skype: patrick_h_lauke
>
>

From: Patrick H. Lauke
Date: Tue, Sep 14 2021 2:21AM
Subject: Re: Does 1.4.10 essentially replace 1.4.4?
← Previous message | No next message

On 14/09/2021 09:09, Mitchell Evan wrote:
[...]
> I agree we should test 1.4.10 in a common desktop display size at 400% (or
> in a dev tools emulator equivalent). The most common aspect ratio of
> desktop screens
> <https://gs.statcounter.com/screen-resolution-stats/desktop/worldwide> is
> 1.78:1 with total usage over 60%, in screens such as 1280x720. Allowing for
> browser and OS toolbars as shown in the Understanding 1.4.10 example video
> <https://www.w3.org/WAI/WCAG21/Understanding/reflow.html#example-1-responsive-design>,
> the
> aspect ratio of the browser viewport (not the screen) is around 2:1.

However, 1.4.10 explicitly codifies one very specific value to test in
terms of width and height. That's the problem with that SC (and yes, it
also doesn't take into account the fact that viewports will indeed be
smaller owing to browser Chrome etc.

>>> 1.4.10 doesn't really say that *both* the width and height must be set.
>
> True — it's a common misunderstanding
> <https://github.com/act-rules/act-rules.github.io/issues/1318> that 1.4.10
> Reflow specifies a 1280x1024 viewport — at most 3% usage, with a very
> different aspect ratio of 1.25:1. The aspect ratio often makes a
> pass-or-fail difference with real-world consequences, when sticky content
> such as a fixed-position footer ends up covering the entire viewport.

Again, the inflexible wording of 1.4.10 cares not for what real-world
sizes of desktops are actually out there, unfortunately. Normatively, it
hardcodes 320 for width and 256 for height, with the explanation/excuse
that this is what you get (though you don't, really) for 1280x1024 at 400%.

It's a borked SC that was rushed without sufficient
scrutiny/forethought...but that's where we are.

P
--
Patrick H. Lauke

https://www.splintered.co.uk/ | https://github.com/patrickhlauke
https://flickr.com/photos/redux/ | https://www.deviantart.com/redux
twitter: @patrick_h_lauke | skype: patrick_h_lauke