Re: Data-* Attributes (Take 2)

Hi Charles,

> Won’t this mean that now we are asking HTML for more than a 1-3 new
attributes?

Sadly, it appears that way. But, if you go back and review Lisa's page
<https://github.com/w3c/personalization-semantics/wiki/Protoypes-with-data-dash>,
there are already 6 proposals there too (including additional variants of
proposals):

*Use with different alternative text*

Proposed attribute 1:  data-alt-numberfree (or
alternatively data-numberfree)


Proposed attribute 2:  data-alt-easylang (or alternatively data-easylang)




*Use with symbol*

Proposed attribute 3: data-map-symbol (or
alternatively data-alternative-symbol, data-purpose-symbol, data-purpose )

*Use with field, destination and action (purpose)*

Proposed attribute 4:  data-field (or alternatively data-purpose-field,
data-purpose)


Proposed attribute 5: data-action

Proposed attribute 6: data-destination

...in fact, the only additions I added were ideas we've capture
previously: *data-literal
*(raining cats and dogs example) and *data-explain* ("press enter to send
the email"). These were either accidentally left off of Lisa's page or this
group had decided to abandon them (during a meeting I might have missed).

In many of the the original examples from Lisa's page, the proposal was an
over-hyphenated proposed attribute (data-alt-numberfree, d
ata-purpose-symbol, data-purpose-field) which I strongly oppose. We were
counselled at TPAC to avoid a micro-syntax by more than one commentor from
Web Platform WG, and if we want to continue to try and work with them, I
will argue we need to also listen to them.

All I have done then is massaged the concepts presented on the original
page, and added the 2 missing attributes (literal and explain). However, I
personally believe my suggestions are 'cleaner' in authoring syntax, and
remain consistent with using the data-* authoring pattern going forward.

Lisa writes:

> I think the most popular will get in. IE the ones that are used in the
real world. So the most important thing is getting things out there to real
world implementors

...and I agree. The advantage of using data-* is that it won't create
validation errors as implementors experiment with the new attributes.

We will learn (over a short to medium timeframe) which attributes are
getting real traction, and those which languish (for whatever reason). Once
we have that body of evidence, we can then go back to Web Plat WG and ask
for (and I'm guessing here) 4 of the 8 to become "full-fledged" HTML 5
attributes, and the balance might then be alternatively prefixed (and at
that time we can have the discussion around using coga-* or aria-* or...
 Note: I am strongly opposed to aui-* because that prefix has already been
taken <https://docs.atlassian.com/aui/> by Atlassian, the makers of Jira
and Confluence (see also here
<https://docs.atlassian.com/aui/0.0.5-9-0-snapshot-035-do-not-use/docs/buttons.html>)
and I have serious concerns about attribute collision.

Additionally, with no offense to Lisa, I found the section on symbols
confusing and unworkable: as I understand it, the intention with symbol is
that it would *replace* on-screen text, and not be added to the page
content by the author (i.e. <img data-alternative-symbol="
http://blisssymbolics.org/refnumber/001" href="mysymb.bmp">) In that
example, the symbol isn't an alternative, it's an embedded image visible to
all (unless the thinking was that the data-alternative-symbol "flag" was
meant to hide the symbol from all, except those that need it - which I find
onerous and heavy-handed).

Finally, while Lisa's examples reference the Bliss collection of symbols, I
personally find it a bit of a stretch that this would be the only symbol
set used on the web - I have to believe that users of these affordances
already have other licensed symbol sets installed, and so the "heavy
lifting" isn't so much choosing the attribute, but rather ensuring that the
symbol sets map to our *token values*, and the transformation is done by
helper apps/plug-ins/etc.

My $0.02 Cdn

JF


On Mon, Apr 1, 2019 at 4:59 PM Charles LaPierre <charlesl@benetech.org>
wrote:

> Hi John,
> I like this approach but I am also worried that if we have  data-literal,
> data-numberfree, data-easylang, data-explain, … how will we go from using
> data- to some other native implementation.  Won’t this mean that now we are
> asking HTML for more than a 1-3 new attributes?  I agree it makes it super
> easy to understand and use but it won’t get adopted by HTML browsers unless
> I am missing something.
>
> Thanks
> EOM
> Charles LaPierre
> Technical Lead, DIAGRAM and Born Accessible
> Twitter: @CLaPierreA11Y
> Skype: charles_lapierre
> Phone: 650-600-3301
>
> On Apr 1, 2019, at 2:53 PM, James Cambron <james.cambron@outlook.com>
> wrote:
>
> Most of what John has included is along the lines with my thinking as
> well. It isn't clear which use cases I added to the original prototype
> document but they were:
>
>
>    - <button data-action="undo">Revert</button>
>    - <a href="home.html" data-destination="homepage">our main page</a>
>
>
> Best,
> Thad
> ------------------------------
> *From:* John Foliot <john.foliot@deque.com>
> *Sent:* Monday, April 1, 2019 2:13 PM
> *To:* public-personalization-tf
> *Subject:* Data-* Attributes (Take 2)
>
> Hi all,
>
> After today's conference call I have some very serious concerns on how
> this is moving forward. Rather than augmenting the current examples, I have
> instead provided an alternative proposal on how to move forward, building
> upon existing consensus (as I understand it).
>
> Please see:
> https://github.com/w3c/personalization-semantics/wiki/Prototypes-with-data-dash-*-(Take-2)
> <https://eur04.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fw3c%2Fpersonalization-semantics%2Fwiki%2FPrototypes-with-data-dash-*-(Take-2)&data=02%7C01%7C%7C157f284c154249f9080508d6b6e6fae8%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636897500497105237&sdata=sKo6jGzNbczIuahE1NJC2%2BkmdWtg%2F3qio6s%2BMpcXsjI%3D&reserved=0>
>
> JF
> --
> *John Foliot* | Principal Accessibility Strategist | W3C AC Representative
> Deque Systems - Accessibility for Good
> deque.com
> <https://eur04.safelinks.protection.outlook.com/?url=http%3A%2F%2Fdeque.com%2F&data=02%7C01%7C%7C157f284c154249f9080508d6b6e6fae8%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C636897500497115245&sdata=1OHNaeCS63OaOpVyIinh8yiMLH0ddNa98tWrXKCQW8U%3D&reserved=0>
>
>
>

-- 
*​John Foliot* | Principal Accessibility Strategist | W3C AC Representative
Deque Systems - Accessibility for Good
deque.com