E-mail List Archives

Search E-mail List Archives for

Number of posts in this thread: 6 (In chronological order)

From: Steve Green
Date: Sat, Sep 02 2023 11:30AM
Subject: WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
No previous message | Next message →

Can anyone help me understand how to apply WCAG SC1.3.5 (Identify Input Purpose) to UK addresses?

We are possibly unusual insofar as our administrative geography is entirely different from that of our postal system.
https://en.wikipedia.org/wiki/Administrative_geography_of_the_United_Kingdom
https://en.wikipedia.org/wiki/List_of_post_towns_in_the_United_Kingdom

In the description of the "autocomplete" values, WCAG says address-level1 is the broadest administrative level in the address, and that in the UK it is the post town. However, many people never use their post town in their address, and may not even know what it is. Historically, post towns were where the delivery centres were, and may be many miles from where you live.

If we follow WCAG and use address-level1 for the post town, which "autocomplete" value do we use for the county? We can't use address-level2 because that's a lower level. And the only higher level is country. And what value do we use for the actual town?

Or should we ignore WCAG and use address-level1 for the county, address-level2 for the town and address-level3 for the district, which is what seems to be the sensible approach?

The whole thing is so ambiguous that different designers will have undoubtedly come to different decisions, undermining the usefulness of the success criterion. The whole problem stems from WCAG saying that address-level1 is the UK post town. If that could be removed or changed to say it's the county, everything would be resolved.

Steve Green
Managing Director
Test Partners Ltd

From: Patrick H. Lauke
Date: Sat, Sep 02 2023 11:45AM
Subject: Re: Fwd:WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
← Previous message | Next message →

Accidentally replied directly, rather than to the list...

P

-------- Forwarded Message --------
Subject: Re: [WebAIM] WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
Date: Sat, 2 Sep 2023 18:44:31 +0100
From: Patrick H. Lauke < = EMAIL ADDRESS REMOVED = >
To: Steve Green < = EMAIL ADDRESS REMOVED = >


On 02/09/2023 18:30, Steve Green wrote:
[...]
> The whole thing is so ambiguous that different designers will have undoubtedly come to different decisions, undermining the usefulness of the success criterion. The whole problem stems from WCAG saying that address-level1 is the UK post town.

WCAG is copying verbatim what's in
https://html.spec.whatwg.org/multipage/form-control-infrastructure.html#autofilling-form-controls:-the-autocomplete-attribute
... so in the first instance, this should be changed there.

> If that could be removed or changed to say it's the county,
everything would be resolved.

I'd suggest actually asking about changes that should be made to WCAG
... in the actual WCAG mailing list or even better the WCAG issue
tracker on github.

P
--
Patrick H. Lauke

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



--
Patrick H. Lauke

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

From: Birkir R. Gunnarsson
Date: Sat, Sep 02 2023 2:58PM
Subject: Re: WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
← Previous message | Next message →

The spirit of 1.3.5 is that user software can autofill the address inputs.
If there is no clear and user agent supported HTML autocomplete
value/set of values for a UK address, i.e. if the user's software
either doesn't understand the value or it's applied inconsistently so
software may get half the fields wrong, then there's limited value in
trying to implement it at this time.
If the user extension either doesn't do anything or, worse yet, fills
in the form wrong, we haven't done the user any favors.

Your best bet would be to test out different values of autocomplete
with things like Google or other popular extension to see if you can
find a set of values that work consistently across browsers.

I'm guessing the requirement is meant to encourage or drive support
for these types of tools. While I like that thought, the biggest
drawback is that it doesn't go well with the unforgiving nature of
WCAG conformance.

On 9/2/23, Steve Green < = EMAIL ADDRESS REMOVED = > wrote:
> Can anyone help me understand how to apply WCAG SC1.3.5 (Identify Input
> Purpose) to UK addresses?
>
> We are possibly unusual insofar as our administrative geography is entirely
> different from that of our postal system.
> https://en.wikipedia.org/wiki/Administrative_geography_of_the_United_Kingdom
> https://en.wikipedia.org/wiki/List_of_post_towns_in_the_United_Kingdom
>
> In the description of the "autocomplete" values, WCAG says address-level1 is
> the broadest administrative level in the address, and that in the UK it is
> the post town. However, many people never use their post town in their
> address, and may not even know what it is. Historically, post towns were
> where the delivery centres were, and may be many miles from where you live.
>
> If we follow WCAG and use address-level1 for the post town, which
> "autocomplete" value do we use for the county? We can't use address-level2
> because that's a lower level. And the only higher level is country. And what
> value do we use for the actual town?
>
> Or should we ignore WCAG and use address-level1 for the county,
> address-level2 for the town and address-level3 for the district, which is
> what seems to be the sensible approach?
>
> The whole thing is so ambiguous that different designers will have
> undoubtedly come to different decisions, undermining the usefulness of the
> success criterion. The whole problem stems from WCAG saying that
> address-level1 is the UK post town. If that could be removed or changed to
> say it's the county, everything would be resolved.
>
> Steve Green
> Managing Director
> Test Partners Ltd
> > > > >


--
Work hard. Have fun. Make history.

From: Steve Green
Date: Sun, Sep 03 2023 3:18AM
Subject: Re: WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
← Previous message | Next message →

According to the authors of this success criterion, the objective has nothing to do with auto-filling forms. It's to allow user agents to reliably understand the purpose of each field so they can use that to help people with cognitive impairments, such as by displaying icons next to them. I don't believe any user agents do that at the moment, but that's not relevant to WCAG conformance.

Reverse engineering Google and other autofill features isn't going to work because my understanding is that they identify fields by the visible text label and they only recognise "autocomplete" values of "on" and "off". They don't use the values specified in WCAG.

Steve

From: Birkir R. Gunnarsson
Date: Sun, Sep 03 2023 6:36AM
Subject: Re: WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
← Previous message | Next message →

Good point, my primary argument still stands, the cart was put a dozen
miles before the horse, I don't see any way to satisfy this criterion
currently, meaning it's impossible to be WCAG level AA conformant.
Needless to say I'd be happy to be wrong about this.

On 9/3/23, Steve Green < = EMAIL ADDRESS REMOVED = > wrote:
> According to the authors of this success criterion, the objective has
> nothing to do with auto-filling forms. It's to allow user agents to reliably
> understand the purpose of each field so they can use that to help people
> with cognitive impairments, such as by displaying icons next to them. I
> don't believe any user agents do that at the moment, but that's not relevant
> to WCAG conformance.
>
> Reverse engineering Google and other autofill features isn't going to work
> because my understanding is that they identify fields by the visible text
> label and they only recognise "autocomplete" values of "on" and "off". They
> don't use the values specified in WCAG.
>
> Steve
>
>
>

From: Patrick H. Lauke
Date: Sun, Sep 03 2023 7:41AM
Subject: Re: WCAG SC1.3.5 (Identify Input Purpose) and UK addresses
← Previous message | No next message

On 03/09/2023 13:36, Birkir R. Gunnarsson wrote:
> Good point, my primary argument still stands, the cart was put a dozen
> miles before the horse, I don't see any way to satisfy this criterion
> currently, meaning it's impossible to be WCAG level AA conformant.
> Needless to say I'd be happy to be wrong about this.

I've been critical of this SC even when it was being devised, for
exactly that reason...chicken and egg/speculative/wishful thinking SC
that has no real-world impact to be honest.

My take has been: do the best you can, choose whichever autocomplete
attribute feels most appropriate to you. Total absence of autocomplete
(or, theoretically, some other way of "programmatically identifying" the
purpose, like RDF tuples...not that they have any real-world support in
anything) is a failure, but unless wildly incorrect, use of an
autocomplete attribute that is arguably (with some subjective
interpretation, as would be the case specifically here for what the top
administrative level is or isn't) "right" passes.

P
--
Patrick H. Lauke

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