Re: Is Common Failure Technique F87: inserting non-decorative content by using ::before and ::after pseudo-elements; still valid?

I was just about to post the same. The idea of progressive enhancement in
the era of evergreen browsers harks to a bygone era. Progressive
enhancement is dead.

Furthermore, it's actually detrimental to accessibility rather than
supportive of it because it forces an architecture that is terrible UX for
everyone.

On Saturday, March 16, 2024, Adam Cooper <cooperad@bigpond.com> wrote:
> Great for hand-coded cottage industry web sites, but, in the real world,
most off-the-shelf front ends are about speed of assembly … foggy windows
into data for which progressive enhancement/graceful degradation is a
quaintness from yesteryear … let’s not forget that progressive enhancement
gave us those horrid to use menus comprising endless nested lists of links
… in my opinion, this kind of nostalgia is what can make accessibility such
a hard sell.
>
>
>
>
>
> From: Steve Green <steve.green@testpartners.co.uk>
> Sent: Sunday, March 17, 2024 9:09 AM
> To: Michael Livesey <mike.j.livesey@gmail.com>
> Cc: David Woolley <forums@david-woolley.me.uk>; w3c-wai-ig@w3.org
> Subject: RE: Is Common Failure Technique F87: inserting non-decorative
content by using ::before and ::after pseudo-elements; still valid?
>
>
>
> You’re missing the point. By using progressive enhancement, you can have
all the latest CSS techniques and have a website that works without CSS. As
the customer, the UK government is entitled to demand you do that when
building their websites. If you’re not prepared to, you don’t get their
business.
>
>
>
> The gov.uk website contains millions of pages, many of which are quite
old, so you will always be able to find bad examples. But the new stuff
(built using the GDS Service Standard I linked to) is really good. I doubt
if any organisation does more user research and user testing. It’s almost
impossible to recruit good accessibility testers in the UK because the
government has recruited hundreds of them and even more UX researchers.
>
>
>
> As for “being accepted by the community”, it’s not for developers to
accept anything other than their responsibility to best serve the end users
and other stakeholders. And end users neither know nor care about the
latest CSS techniques.
>
>
>
> Steve
>
>
>
> From: Michael Livesey <mike.j.livesey@gmail.com>
> Sent: Saturday, March 16, 2024 7:09 PM
> To: Steve Green <steve.green@testpartners.co.uk>
> Cc: David Woolley <forums@david-woolley.me.uk>; w3c-wai-ig@w3.org
> Subject: Re: Is Common Failure Technique F87: inserting non-decorative
content by using ::before and ::after pseudo-elements; still valid?
>
>
>
> We will have to agree to disagree, Steve.
>
> Anyone who has done a UK tax return or tried to set up a childcare tax
account knows that .gov sites are notoriously bad UX.
>
> Regardless, as I said above, F87 is illogical because there will be more
issues than just ::before or ::after content problems for most websites.
Why focus on those two pseudo elements?
>
> The vast majority of websites are a complete mess without CSS and fail
dozens of criteria. If the intent is to make websites accessible without
CSS, it would have the effect that most modern CSS techniques would not be
able to be used and that is something that would not be accepted by the
community.
>
>
>
>
>
> On Saturday, March 16, 2024, Steve Green <steve.green@testpartners.co.uk>
wrote:
>> Actually, the opposite is true. In user testing, UK government websites
perform better than most other websites. People literally ask why all
websites can’t be like that. Of course some of the old parts of the website
aren’t as good, but the usability and accessibility of the newer parts are
mostly excellent. The design is rather boring, but it’s mostly designers
who obsess over that – users really don’t care.
>>
>>
>>
>> And disabling CSS doesn’t need to cause problems with focus, resize,
reflow etc. In fact, those things work perfectly without CSS. If websites
don’t work without CSS, it’s because most developers don’t learn how to do
progressive enhancement these days. But 20 years ago progressive
enhancement was a common practice and websites often worked fine without
CSS and JavaScript. They had to in order to meet WCAG 1.0. And it would
still be perfectly possible to do that with most websites today.
>>
>>
>>
>> You can still use modern selectors such as :not or :has or
:focus-within. You just need to design the website to work without them,
then add them to give the appearance and behaviour you want. The appearance
and behaviour without them might not be as slick, but there’s no reason why
the website couldn’t still be usable.
>>
>>
>>
>> Steve
>>
>>
>>
>> From: Michael Livesey <mike.j.livesey@gmail.com>
>> Sent: Saturday, March 16, 2024 7:21 AM
>> To: David Woolley <forums@david-woolley.me.uk>
>> Cc: w3c-wai-ig@w3.org
>> Subject: Re: Is Common Failure Technique F87: inserting non-decorative
content by using ::before and ::after pseudo-elements; still valid?
>>
>>
>>
>> I think Phil has a valid point. If one disconnected all CSS, most modern
web pages would have more accessibility failures than just :before and
:after elements, so why single out those two pseudo-elements?
>>
>> Modern web pages are infinitely more complex than was envisaged in 1999
or 2015. Disabling CSS would cause focus, resize, reflow, use of colour and
goodness knows how many other failures. It would make using modern
selectors such as :not or :has or :focus-within impossible to use.
>>
>> Kudos for the UK government insisting on web pages work without CSS,
although anyone who has visited a .gov website will agree, they are
terrible UX and difficult for non-accessible users to navigate let alone
accessible users. They should be held more as an example of why well
designed CSS should be required.
>>
>> Perhaps it is time F87 was removed?
>>
>>
>>
>> On Friday, March 15, 2024, David Woolley <forums@david-woolley.me.uk>
wrote:
>>> On 15/03/2024 22:10, Phill Jenkins wrote:
>>>>
>>>> temporary network errors [hmm, network errors impact HTML same as CSS]
>>>
>>> I still encounter these, and they are more disconcerting than a
complete failure to load, as would be the case for HTML.
>>>
>>>