Re: FYI: proposed ARIA feature: synonyms

Thank you for the clarification.
It is an understandable convention, given that most HTML attributes consist
of 1 or 2 English words (or abbreviations thereof), so the compact forms
have low cost, if any.
Yet when the names grow to 3-4 words, especially if not monosyllabic, there
is clear loss in legibility.

As for MathML, I have also admired "scriptsizemultiplier" and
"superscriptshift", although they are rather exceptional cases.

I suppose I can see the glass as half full and be grateful that the
discussion is not about an attribute named "arialabelsynonyms" and that
MathML does not need to employ a namespaced "math-scriptsizemultiplier".

Apologies for the off-topic comments,
Deyan


On Mon, Feb 6, 2023 at 7:32 PM Neil Soiffer <soiffer@alum.mit.edu> wrote:

> FYI: "aria-" is used as a pseudo namespace. They don't use "-" in any of
> their names. That is consistent with the rest of HTML (5) attribute names
> such as "contextmenu", "readonly", "tabindex", etc. It's true for (most?
> all?) MathML (e.g., "maxsize", "rowspacing"). That's just convention
> because it is legal to use "-".
>
> FYI 2: Although the issue I pointed to is old, if you look at the end of
> the issue, you'll see a link to a PR and that has recently had a bunch of
> discussion. That's what brought it to my attention.
>
>     Neil
>
>
> On Mon, Feb 6, 2023 at 4:12 PM Deyan Ginev <deyan.ginev@gmail.com> wrote:
>
>> Hi Neil,
>>
>> Thank you for bringing the aria-labelsynonyms attribute to our attention,
>> it was the first time I learned of its proposal.
>>
>> Two notes come to mind:
>> 1.  my "alias" suggestion had the additional benefit of it being a
>> one-time global annotation (in the Open and possibly Core lists), rather
>> than being locally generated on every <math> element that may need it.
>> 2. I don't really understand the aesthetics of half-dashed naming
>> schemes. "aria-label-synonyms" feels much cleaner/legible than
>> "aria-labelsynonyms".
>>     I assume that is some internal consistency ARIA is maintaining,
>> considering e.g. "aria-brailleroledescription" suffers the same legibility
>> deficit.
>>
>> For readers not familiar with the "alias" discussion, the issue is at
>> https://github.com/w3c/mathml-docs/issues/40
>>
>> Greetings,
>> Deyan
>>
>>
>> On Mon, Feb 6, 2023 at 6:50 PM Neil Soiffer <soiffer@alum.mit.edu> wrote:
>>
>>> The ARIA WG is considering adding aria-labelsynonyms
>>> <https://github.com/w3c/aria/issues/1038>. I won't pretend to
>>> understand the subtleties. The gist is that for speech-to-text (not
>>> text-to-speech), adding synonyms makes it more likely that something will
>>> be found. There is a little bit of an analogy with Deyan's goal of adding
>>> aliases for intent names. Perhaps others will not feel there is much
>>> analogy or maybe someone will find a strong similarity. Given the potential
>>> similarities, I felt people should be aware of it.
>>>
>>>     Neil
>>>
>>>
>>>