Re: [css3-speech] Drop phonemes section, mark properties

On 10/20/2010 04:33 PM, Belov, Charles wrote:
>fantasai wrote:
>> Just from a quick scan through the last WD...
>>
>> http://www.w3.org/TR/2004/WD-css3-speech-20041216/#phonetic-props
>>
>> The phonemes section of css3-speech seems a little too
>> closely tied to the content to be stylistic. I suggest
>> dropping it; the use case should be handled by the markup somehow.
>
> Would you be so kind as to provide an example as to how this would be
> handled by the markup?

The equivalent thing could be done in markup with e.g. a 'phonemes'
attribute rather than mapping it to an ID and then to a style sheet.
But I believe SSML addresses this use case already.

> @phonetic-alphabet "ipa";
> ..tomato { phonemes: "t\0252 m\0251 to\028a " }
>
> unless one is sure there will be only one occurrence of the word
> "tomato" on the page).

Right, and

> Better yet would be a content-based selector:
>
> @phonetic-alphabet "ipa";
> [\wtomato\w] { phonemes: "t\0252 m\0251 to\028a " }
>
> which would leave the markup clean:

By which you mean a pseudo-element, and no, there are no plans to
add such a thing to CSS. There are only plans not to add it to CSS.

> I do have a real-world concern use case that "Muni" (the San Francisco
> public transportation system) is pronounced Mew-nee not Moo-nee, and it
> would be tedious/bloating to have to tag it wherever it occurs rather
> than being able to specify it once in a style sheet.  This also applies
> to many street names in San Francisco, and, I imagine elsewhere, as well
> as many company names.
>
> Or perhaps there needs to be a vocabulary-pronunciation download
> facility similar to what is being done with downloadable fonts.

and here we are. The feature in CSS3 Speech is neither well-placed
in a style sheet, nor very efficient.

~fantasai

Received on Thursday, 21 October 2010 00:02:22 UTC