W3C home > Mailing lists > Public > www-style@w3.org > February 2011

Re: "phonemes" property in the CSS3 Speech module

From: Daniel Weck <daniel.weck@gmail.com>
Date: Fri, 4 Feb 2011 10:32:11 +0000
Message-Id: <4972EB73-F5A2-4188-BA84-615FE011DA1B@gmail.com>
To: "www-style@w3.org list" <www-style@w3.org>
Hi Christoph, I don't think that the goal is to CSS-ize W3C PLS (or  
any other existing pronunciation lexicon format), which is what your  
proposal seems to be doing. Furthermore, using language codes only  
addresses particular use-cases, such as accents and dialects (see my  
previous email where I illustrate the point, using the word "read" in  
british english).
Regards, Daniel

On 4 Feb 2011, at 10:12, Christoph Päper wrote:

> I believe (custom) language codes are the best solution for  
> pronunciation hints. This is somewhat similar to hyphenation.
>
>  <html lang="en">
>  <body lang="en-EU">
>    <dialog lang="en-GB">
>    <dt>Gavin</dt><dd lang="en-AU">My doctor said to eat
>      a <span lang="en-US">tomato</span> every day.</dd>
>    <dt>Colm</dt><dd lang="en-IR"><em>My</em> doctor said to take
>      <em>my <span lang="en-GB">vitamin</span></em> pills every day.</ 
> dd>
>    </dialog>
>  </body>
>  </html>
>
>  @pronunciation "en" {src: url("/dict/en.prc");}
>  @pronunciation "en-US", "en" {word-forms:
>    tomato = "təˈmeɪtoʊ" /
>    vitamin = "ˈvaɪtamɪn";
>  }
>  @pronunciation "en-GB", "en-AU", "en-EU" {word-forms:
>    tomato = ipa(təˈmɑːtəʊ), sampa(t@"mA:t@U);
>  }
>  @pronunciation "en-GB", "en-EU" {word-forms:
>    vitamin = "ˈvɪt.ə.mɪn";
>  }
>  @pronunciation "en-AU" {word-forms:
>    vitamin = "ˈvaɪ.tə.mən";
>  }
>  @pronunciation "en-CA" {word-forms:
>    tomato = "təˈmɛɪtoː", "təˈmætoː";
>    /* two values are legal, but the second one is unused */
>  }
>  @pronunciation "en-X-nuspik" {word-forms:
>    tomato = "ˈtɔməto" /
>    vitamin = "ˈvɪˈtaːmɪn";
>  }
>
>  @hyphenation "de" {src: url("/dict/de.spl");
>    word-forms: "Pä-per" /* not in dictionary, but common to all  
> variants */;
>  }
>  @hyphenation "de-1900", "de-Latf" {word-forms:
>    "Decke" = "Dek-ke" / /* mandatory ‘ck’ ligature in Fraktur  
> split */
>    "Liste" = "Li-ste" / /* mandatory ‘st’ ligature in Fraktur  
> kept,
>    automatically includes long-s ‘ſ’ variant "Li-ſte" */
>    /* using short form from here on */
>    "Ma-gnet" / /* etymologic = morphologic, not phonemic */
>    "Ma-gne-te";
>  }
>  @hyphenation "de-1996", "de-Latn" {word-forms:
>    "De-cke" / "Lis-te" /
>    "Mag-net", "Ma-gnet" / "Mag-ne-te", "Ma-gne-te"
>  /*^ either of the options is valid, but only one of them must be  
> used */;
>  }
>  @hyphenation "de-X-Paeper" {word-forms:
>    "Dek-ke" / "Lis-te" / "Mag-net" / "Mag-ne-te";
>  }
>
> Due to time constrains I probably didn’t follow BCP47 and current  
> CSS drafts correctly, which I should have done. IPA and SAMPA  
> pronunciations were taken from Wiktionary.
>
> U+2027 “Hyphenation Point” ‘·’ or U+007C “Vertical Line”  
> ‘|’ could be used instead of U+002D “Hyphen-Minus” ‘-’ to  
> indicate hyphenation opportunities, since the hyphen may occur  
> inside words naturally. ‘word-forms’ should be a last resort  
> property, ideally other properties are able to define regular  
> hyphenation patterns.
Received on Friday, 4 February 2011 10:32:47 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:37 GMT