- From: Gregg Vanderheiden <gv@trace.wisc.edu>
- Date: Mon, 6 Mar 2006 11:14:01 -0600
- To: <public-wcag-teamb@w3.org>
JUST FYI since John is offline today and you are working on this. Gregg -- ------------------------------ -----Original Message----- From: WATANABE Takayuki [mailto:nabe@lab.twcu.ac.jp] Sent: Monday, March 06, 2006 10:39 AM To: Gregg Vanderheiden Cc: 'Makoto UEKI - Infoaxia, Inc. -'; 'John M Slatin'; 'Ben Caldwell' Subject: Re: JIS comments and Pronunciation issue Dear Gregg, I think the current SC is good. I agree we do the rest with how to meet documents. --Takayuki On Mon, 6 Mar 2006 10:10:38 -0600 "Gregg Vanderheiden" <gv@trace.wisc.edu> wrote: > Thank you Takayuki, > > Do you think the current SC is good enough? Can we do the rest with how we > handle the 'how to meet' document and techniques? > > > > Gregg > > -- ------------------------------ > Gregg C Vanderheiden Ph.D. > Professor - Ind. Engr. & BioMed Engr. > Director - Trace R & D Center > University of Wisconsin-Madison > The Player for my DSS sound file is at http://tinyurl.com/dho6b > > -----Original Message----- > From: WATANABE Takayuki [mailto:nabe@lab.twcu.ac.jp] > Sent: Monday, March 06, 2006 7:27 AM > To: Gregg Vanderheiden > Cc: 'Makoto UEKI - Infoaxia, Inc. -'; 'John M Slatin'; 'Ben Caldwell' > Subject: Re: JIS comments and Pronunciation issue > > Dear Gregg, > > Thank you very much for your response. > > On Sun, 5 Mar 2006 16:26:00 -0600 > "Gregg Vanderheiden" <gv@trace.wisc.edu> wrote: > > > - with Han - you have to mark up many or perhaps all. > > - with Han and proper names - you must mark up all. > > - with hiragana (kana) you should not have to mark things up since > > it is phonetic. But is sounds like you do sometimes anyway. Is this correct? > > > Is that good enough for your needs? > > I think it's too strict. :-) > I'm sorry my previous explanation was not good enough. > > > A Japanese screen reader has its own dictionary which probably have > several thousands key-value pairs of a single character and its > pronunciation. Most of the entry is Han characters. > Some entries are words which pronunciation is difficult. > A text-to-speech engine also has its own dictionary. > With use of these dictionaries Japanese screen readers read aloud text well. > But sometime it cannot read it correctly. Failures occur when > > 1) There is not an entry in dictionaries. > Example of this case is uncommon proper nouns. > > 2) There are multiple entries of the same word or character. > There are a few Han words or characters which pronunciation cannot be > determined by syntactics. Semantics or pragmatics also do not help. > Only an author can decide its pronunciation. > > Thus, I should say > - You do not have to mark phonetic characters such as hiragana and > katakana (kana). > > - You should mark up a Han character or a word which includes Han > characters if it is uncommon to the intended users. Uncommon names > of persons or places, and very difficult and uncommon Han character are > good examples of these cases. > > - You should mark up some Han characters or some words which includes > Han characters which have multiple pronunciation (reading). > You do not have to mark up all these cases because in most cases screen > reading software can determine the correct reading. > > I admit it is ambiguous description because it depends on the > capability of screen reading software. > Perhaps we need a concept of baseline screen reading software. > I think we need heuristics approach as well as algorithm one. > > > > The way I wrote the guideline - you would have to use markup > > whenever the author knew that the word would be ambiguous. > > Yes. > > > This would cover all HAN > > proper names and any HAN characters that would always be ambiguous > > in a sentence. > > Yes. > > > It would NOT cover things that are obvious from the sentence, but > > the screen reader is just not very good and will mispronounce it. > > Yes. > > > Since the authors > > don't use screen readers (and certainly don't use all possible > > screen > > readers) it would not be possible for an author to know which words > > each different screen reader had trouble with. They could tell > > which words a human would have trouble with and mark up those. > > Perhaps we need a concept of baseline screen reading software. > I think we need heuristics approach as well as algorithm one. > > --Takayuki > > > On Sun, 5 Mar 2006 16:26:00 -0600 > "Gregg Vanderheiden" <gv@trace.wisc.edu> wrote: > > > Hi Takayuki, > > > > Your descriptions is exactly as I understood the problem. > > > > - with Han - you have to mark up many or perhaps all. > > - with Han and proper names - you must mark up all. > > - with hiragana (kana) you should not have to mark things up since > > it is phonetic. But is sounds like you do sometimes anyway. Is this correct? > > > > The way I wrote the guideline - you would have to use markup > > whenever the author knew that the word would be ambiguous. This > > would cover all HAN proper names and any HAN characters that would > > always be ambiguous in a sentence. > > > > It would NOT cover things that are obvious from the sentence, but > > the screen reader is just not very good and will mispronounce it. > > Since the authors don't use screen readers (and certainly don't use > > all possible screen > > readers) it would not be possible for an author to know which words > > each different screen reader had trouble with. They could tell > > which words a human would have trouble with and mark up those. > > > > > > Is that good enough for your needs? > > > > -----Original Message----- > > From: WATANABE Takayuki [mailto:nabe@lab.twcu.ac.jp] > > Sent: Saturday, March 04, 2006 3:14 AM > > To: Gregg Vanderheiden > > Cc: 'Makoto UEKI - Infoaxia, Inc. -'; 'John M Slatin'; 'Ben Caldwell' > > Subject: Re: JIS comments and Pronunciation issue > > > > Dear Gregg, > > > > Thank you very much for your comment. > > Please let me explain our intention of the pronunciation problem: > > > > We use a language to exchange ideas and facts. > > One language can be represented aurally or visually. > > I think the problem is, especially in Han characters, there is no > > accurate matching between aural representation (spoken language) and > > visual representation (written language). > > > > I read a written sentence character by character. Of course I use > > word and context to understand the meaning of the sentence but the > > basic unit of visual perception is a character. > > > > When hearing a sentence I use a phoneme instead of a character. I > > do not care about if a sentence is written in Kanji or kana. > > > > Japanese hiragana (kana) is a sylabogram (I do not know the exact word). > > Each hiragana represents a certain phoneme (sound). > > Kanji (Han characters) are quite different. One Kanji can be read > > differently. Different pronunciations of the same Kanji-words > > represent different meanings. > > > > > > A screen reader reads aloud a written language but it does not read > > character by character. It generally reads by sentence as if that > > sentence is spoken. > > > > In most cases a test-to-speech engine analyzes sentences well but > > some cases a text-to-speech engine cannot know the aural > > representation of some > words. > > Proper nouns such as persons name and name of city or town are good > > examples of such cases. > > > > Sighted uses can read a sentence but the users without vision cannot > > read the sentence. They relies on a screen readers or > > text-to-speech > engine. > > That's why we think this is accessibility issue. > > > > > > As for your comment: > > > > > Maybe I am misunderstanding - but what it looks like you are > > > proposing is true of every screen reader -- even with ordinary > > > words if they are not explicitly coded. So it would require > > > markup on all > pages. > > > > Logically, every content should have pronunciation markups. > > But fortunately, in most cases, screen reading software can analyse > > content to read aloud correctly. > > > > > How would an author know which words a particular persons screen > > > reader won't pronounce correctly? > > > > We cannot define which words should have pronunciation information > > (markup) because we cannot list up every proper nouns. > > But we can show examples. > > > > > Also - it would require markup on pages where the meaning is easy > > > to determine from context but the screen reader just doesn't do it. > > > > In this case, the "reading" is not determined from context. > > > > > > I'm not sure if my comment answers your comment but I hope my > > comment helps to improve Success Criterion 3.1.7. > > > > > > On Fri, 3 Mar 2006 12:21:19 -0600 > > "Gregg Vanderheiden" <gv@trace.wisc.edu> wrote: > > > > > Hi Makoto, > > > > > > Maybe I am misunderstanding - but what it looks like you are > > > proposing is true of every screen reader -- even with ordinary > > > words if they are not explicitly coded. So it would require > > > markup on all > pages. > > > > > > Also - it would require markup on pages where the meaning is easy > > > to determine from context but the screen reader just doesn't do it. > > > > > > This is a screen reader issue rather than in accessible content. > > > > > > How would an author know which words a particular persons screen > > > reader won't pronounce correctly? > > > > > > > > > > > > Gregg > > > > > > -- ------------------------------ Gregg C Vanderheiden Ph.D. > > > Professor - Ind. Engr. & BioMed Engr. > > > Director - Trace R & D Center > > > University of Wisconsin-Madison > > > The Player for my DSS sound file is at http://tinyurl.com/dho6b > > > > > > -----Original Message----- > > > From: Makoto UEKI - Infoaxia, Inc. - [mailto:ueki@infoaxia.co.jp] > > > Sent: Thursday, March 02, 2006 1:24 AM > > > To: John M Slatin > > > Cc: Gregg Vanderheiden; Ben Caldwell; WATANABE Takayuki > > > Subject: Re: JIS comments and Pronunciation issue > > > > > > Hi John, > > > > > > Thanks for the updates. I would like to hear from Takayuki, but > > > the folowings are my concerns. > > > > > > The last sentences in the Intent section: > > > <blockquote> > > > Each pronunciation has a different meaning. Screen readers will > > > not always be able to determine the correct pronunciation. > > > </blockquote> > > > > > > In my opinion, for Japanese, the nuance has to be revised a bit. > > > > > > <SUGGEST> > > > Screen readers may read aloud the characters incorrectly without > > > the information on the pronunciation. When read incorrectly, it > > > will not make sense to the users. > > > </SUGGEST> > > > > > > What I want to say is that different pronunciation/reading won't > > > change > > the > > > meaning it self in genaral. It just won't make sense. In other > > > words, the words pronounced incorrectly doesn't exist. The users > > > will be at a loss because they don't know what the word is. This > > > is the frequent problems in the Japanese content. > > > > > > The rest of the updated draft looks good to me. John, thank you > > > very much for everything you've done for this SC. > > > > > > > > > Cheers, > > > Makoto > > > > > > > > > On Wed, 1 Mar 2006 15:36:21 -0600 > > > "John M Slatin" <john_slatin@austin.utexas.edu> wrote: > > > > > > > Dear Watanabe, Makoto, and Gregg: > > > > > > > > Thanks very much for your comments on the draft I sent yesterday. > > > > Here is another update. > > > > > > > > I've made the following changes: > > > > > > > > - Incorporated Gregg's suggested wording for the Intent section. > > > > - Under Techniques, made the item "Providing pronunciation using > > > > technology-specific techniques" no longer a link (as Gregg > > > > suggested) > > > > - Changed the order of the examples to match Watanabe's > > > > suggested order > > > > > > > > I did not change the wording of the 5th example (about the > > > > content that includes quotations in Chinese and Korean). I think > > > > it is OK not to specify a particular technique. However, if > > > > there is a particular technique that is most common in Japan > > > > (for example, using Ruby markup or providing the pronunciation > > > > information in > > > > parentheses) I will add it to the example. > > > > > > > > Thanks! > > > > > > > > John > > > > > > > > > > > > > > > > > > > > "Good design is accessible design." > > > > John Slatin, Ph.D. > > > > Director, Accessibility Institute University of Texas at Austin > > > > FAC 248C > > > > 1 University Station G9600 > > > > Austin, TX 78712 > > > > ph 512-495-4288, f 512-495-4524 > > > > email jslatin@mail.utexas.edu > > > > web http://www.utexas.edu/research/accessibility/ > > > > > > > > > > > > > > > > > > > > > > > > -----Original Message----- > > > > From: Gregg Vanderheiden [mailto:gv@trace.wisc.edu] > > > > Sent: Tuesday, February 28, 2006 11:50 pm > > > > To: John M Slatin; 'Makoto UEKI - Infoaxia, Inc. -' > > > > Cc: 'Ben Caldwell'; 'WATANABE Takayuki' > > > > Subject: RE: JIS comments and Pronunciation issue > > > > > > > > John > > > > > > > > The INTENT still talks about things that are not part of the > > > > Success criterion. > > > > > > > > Paragraph 1 now reads > > > > "This success criterion helps people who are blind, people who > > > > have low vision, and people with reading disabilities as well as > > > > people without disabilities while also allowing authors to > > > > publish difficult Web content. > > > > Difficult text may be appropriate for most members of the > > > > intended audience (that is, most of the people for whom the > > > > content has been created). But even among highly educated users > > > > with specialized knowledge of the subject matter , there are > > > > often people with disabilities (and people without > > > > disabilities) who may not be able to determine the correct > > > > pronunciation of a word. > > > > > > > > There is nothing in the success criterion about helping people > pronounce > > > > words unless the pronunciation is required for understanding. This > is > > > > rarely the case with complex text. You may have no idea how to > > > > pronounce something but the pronunciation is still unique and > > > > therefore the meaning does not change depending on the pronunciation. > > > > > > > > > > > > Perhaps something like: > > > > > > > > Often words or characters have different meanings each with its own > > > > pronunciation. Usually the meaning can be determined from the > context > > > > of > > > > the sentence. However, for more complex or ambiguous sentences, > > > > or for some languages, the meaning of the word cannot be easily > > > > determined or determined at all without knowing the pronunciation. > > > > When the sentence is read aloud and the screen reader reads the > > > > word using the wrong pronunciation, it can be even more > > > > difficult to understand than when read visually. When words are > > > > ambiguous or indeterminate without knowing the pronunciation is > > > > then providing some means of determining the pronunciation is needed. > > > > > > > > Additionally, in some languages certain characters can be > > > > pronounced in different ways. In Japanese, for example, there > > > > are characters like Han > > > > characters(Kanji) which have multiple pronunciations. Each > > > > pronunciation has a different meaning. Screen readers will not > > > > always be able to determine the correct pronunciation. > > > > > > > > > > > > Also the following item in the sufficient list should not be a > > > > link I don't think. > > > > Providing pronunciation using a technology-specific technique > > > > below > > > > > > > > > > > > Under technology specific techniques - HTML we have the following. > > > > > > > > # Linking to pronunciations . > > > > # Linking to a Glossary Page that includes specific > > > > pronunciation information for each run of text whose meaning > > > > depends on > pronunciation. > > > > # Using standard diacritical marks that can be turned off > > > > > > > > These sound like general techniques. > > > > > > > > If they use a specific HTML construct then that should be listed. > > > > If there are several under each then I would say # Linking to > > > > pronunciations in HTML. > > > > > > > > > > > > Thanks for pulling this together > > > > > > > > > > > > Gregg > > > > > > > > -- ------------------------------ Gregg C Vanderheiden Ph.D. > > > > Professor - Ind. Engr. & BioMed Engr. > > > > Director - Trace R & D Center > > > > University of Wisconsin-Madison > > > > The Player for my DSS sound file is at http://tinyurl.com/dho6b > > > > > > > > -----Original Message----- > > > > From: John M Slatin [mailto:john_slatin@austin.utexas.edu] > > > > Sent: Tuesday, February 28, 2006 3:58 PM > > > > To: Makoto UEKI - Infoaxia, Inc. - > > > > Cc: Gregg Vanderheiden; Ben Caldwell; WATANABE Takayuki > > > > Subject: RE: JIS comments and Pronunciation issue > > > > > > > > Dear Makoto, Takayuki, Gregg, and Ben: > > > > > > > > Here is an updated version of the proposed SC 3.1.7. (The > > > > filename includes today's date, 28February2006.) It incorporates > > > > comments and suggestions from Gregg and Makoto. > > > > > > > > - I deleted the sentence that said, "People with reading > > > > disabilities may need help..." from the first paragraph of the > > > > Intent > section. > > > > - I changed the titles of the techniques about using a glossary; > > > > I hope the new titles accurately reflect Gregg's concerns while > > > > showing more clearly when these techniques can be used. > > > > - I changed the first example. The phrase "words of foreign origin" > > > > does not appear in the "title" now. Instead it talks about > > > > characters that are shared by several languages but pronounced > > > > differently in each language. The example uses the idea of a Web > > > > page on a Japanese university site that includes short > > > > quotations from scholarly works in Chinese and Korean. I hope > > > > this example is > accurate. > > > > Thanks to both Makoto and Gregg for comments and suggestions! > > > > > > > > John > > > > capture > > > > > > > > > > > > > > > > "Good design is accessible design." > > > > John Slatin, Ph.D. > > > > Director, Accessibility Institute University of Texas at Austin > > > > FAC 248C > > > > 1 University Station G9600 > > > > Austin, TX 78712 > > > > ph 512-495-4288, f 512-495-4524 > > > > email jslatin@mail.utexas.edu > > > > web http://www.utexas.edu/research/accessibility/ > > > > > > > > > > > > > > > > > > > > > > > > -----Original Message----- > > > > From: Makoto UEKI - Infoaxia, Inc. - > > > > [mailto:ueki@infoaxia.co.jp] > > > > Sent: Tuesday, February 28, 2006 2:31 pm > > > > To: John M Slatin > > > > Cc: Gregg Vanderheiden; Ben Caldwell; WATANABE Takayuki > > > > Subject: Re: JIS comments and Pronunciation issue > > > > > > > > Hi John, > > > > > > > > Your suggested example addressed my concern. Yes, this is an > > > > issue for all Japanese. And the Japanese screen readers may read > > > > Chinese and Korean names incorrectly. Also people with reading > > > > disabilities may have difficulty to determine the reading of > > > > their > names. > > > > > > > > > > > > Cheers, > > > > Makoto > > > > > > > > On Tue, 28 Feb 2006 14:19:54 -0600 "John M Slatin" > > > > <john_slatin@austin.utexas.edu> wrote: > > > > > > > > > Gregg suggested deleting the following example: > > > > > > > > > > <proposedExample> > > > > > > Web content includes some words of foreign origin. The > > > > > > pronunciation of each word is provided in parentheses right > > > > > > after the word. Screen readers and users who don't know how > > > > > > to pronounce the word can use the pronunciation information > > > > > > to read the word > > > > correctly. > > > > > </blockquote> > > > > > </proposedExample> > > > > > > > > > > But Makoto describes specific circumstances where the > > > > > pronunciation information is needed: > > > > > <blockquote> > > > > > > > > > > MU > > > > > This may be applied to Chinese or Korean name shown in the > > > > > Japanese content. We can use Kanji for their name. But the > > > > > reading/pronunciation for some characters is different from > > > > > Japanese reading. So we need to provide information on > reading/pronunciation. > > > > > </blockquote> > > > > > > > > > > We could change the example so that it's closer to what Makoto > > > > > describes: > > > > > > > > > > <suggestion> > > > > > A Japanese university Web site includes the names of several > > > > > Chinese and Korean scholars. The names are written using the > > > > > same script as the Japanese text. Pronunciation information is > > > > > needed to determine the correct reading of the characters. > > > > > Pronunciation information is provided for each Chinese and > > > > > Korean > name. > > > > > </suggestion> > > > > > > > > > > Additional comment/question: > > > > > Makoto, this sounds like something that is needed for all > > > > > Japanese readers, not just readers with disabilities. Am I > > > > > correct? Is there a specific problem here for users with > disabilities? > > > > > > > > > > John > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > "Good design is accessible design." > > > > > John Slatin, Ph.D. > > > > > Director, Accessibility Institute University of Texas at > > > > > Austin FAC 248C > > > > > 1 University Station G9600 > > > > > Austin, TX 78712 > > > > > ph 512-495-4288, f 512-495-4524 email jslatin@mail.utexas.edu > > > > > web http://www.utexas.edu/research/accessibility/ > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > -----Original Message----- > > > > > From: Makoto UEKI - Infoaxia, Inc. - > > > > > [mailto:ueki@infoaxia.co.jp] > > > > > Sent: Tuesday, February 28, 2006 1:37 pm > > > > > To: Gregg Vanderheiden > > > > > Cc: John M Slatin; 'Ben Caldwell'; WATANABE Takayuki > > > > > Subject: Re: JIS comments and Pronunciation issue > > > > > > > > > > Hi John, > > > > > > > > > > Here are my comments. My comments begin with MU below. > > > > > > > > > > <blockquote> > > > > > > PROPOSED > > > > > > People with reading disabilities may also need help in > > > > > > associating the > > > > > > > > > > > text with the correct pronunciation. > > > > > > > > > > > > COMMENT > > > > > > I don't understand this last sentence. I think it is covered > with > > > > > the text > > > > > > above and should be deleted. If it has another meaning not > > > > > > covered above then it needs to be re-written. Also check to > > > > > > be sure that the SC actually meets the need if it is a new > > > > > > need not covered > > > > above. > > > > > </blockquote> > > > > > > > > > > MU > > > > > I think it is covered by Gregg's suggestion. So we can delete > > > > > this sentence. > > > > > > > > > > <blockquote> > > > > > > PROPOSED > > > > > > Techniques for Addressing Success Criterion 3.1.7 > > > > > > > > > > > > The following combinations of techniques are deemed to be > > > > > > sufficient > > > > > > > > > > by the WCAG Working Group for meeting success criterion 3.1.7. > > > > > > > > > > > > Provide the pronunciations of words using at least one of > > > > > > the techniques from the following list or one of the > > > > > > technology-specific > > > > > techniques. > > > > > > > > > > > > * Providing pronunciations in parentheses. > > > > > > * Linking to pronunciations . > > > > > > * Using a Glossary Page that includes pronunciation > information. > > > > > > * Providing pronunciation using a technology-specific > > > > > > technique below > > > > > > > > > > > > COMMENT > > > > > > The third bullet is not sufficient unless there is only one > > > > > > unique meaning and pronunciation for the word used in the > document. > > > > > > > > > > > > SUGGEST > > > > > > Change the list to > > > > > > > > > > > > * Providing pronunciations in parentheses. > > > > > > * Linking to specific pronunciations. > > > > > > * Providing the correct pronunciation using a > > > > > > technology-specific > > > > > > > > > > > technique below > > > > > > * If (and only if) the word has only one pronunciation in > > > > > > the document, using a Glossary Page that includes > > > > pronunciation > > > > > > information. > > > > > </blockquote> > > > > > > > > > > MU > > > > > I couldn't understand the situation where Gregg made change. > > > > > Need more > > > > > > > > > description for this. > > > > > > > > > > <blockquote> > > > > > > PROPOSED > > > > > > Technology-Specific Techniques HTML Techniques > > > > > > > > > > > > * Using the ruby element (XHTML 1.1). > > > > > > * Linking to pronunciations. > > > > > > * Linking to a Glossary Page that includes pronunciation > > > > > information. > > > > > > * Using standard diacritical marks that can be turned > > > > > > off > > > > > > > > > > > > COMMENT > > > > > > Bullet 3 doesn't work unless only one pronunciation is provided. > > > > > > > > > > > > SUGGESTION > > > > > > Change bullets to > > > > > > * Using the ruby element (XHTML 1.1). > > > > > > * Linking to specific pronunciations. > > > > > > * Linking to a Glossary Page that includes pronunciation > > > > > > information if only one pronunciation is provided in Glossary. > > > > > > * Using standard diacritical marks that can be turned > > > > > > off > > > > > </blockquote> > > > > > > > > > > MU > > > > > Same as above. > > > > > > > > > > <blockquote> > > > > > > THE REST LOOKS FINE EXCEPT FOR THE FOLLOWING BENEFIT -- > > > > > > WHICH IS DIFFENT FROM THE SUCCESS CRITIERION. PRONUNCIATION > > > > > > NOT REQUIRED FOR > > > > > FOREIGN WORDS. > > > > > > SUGGEST DELETING > > > > > > > > > > > > * Text that includes pronunciation information for a > > > > > > word of foreign origin. > > > > > > Web content includes some words of foreign origin. The > > > > > > pronunciation of each word is provided in parentheses right > > > > > > after the word. Screen readers and users who don't know how > > > > > > to pronounce the word can use the pronunciation information > > > > > > to read the word > > > > correctly. > > > > > </blockquote> > > > > > > > > > > MU > > > > > This may be applied to Chinese or Korean name shown in the > > > > > Japanese content. We can use Kanji for their name. But the > > > > > reading/pronunciation for some characters is different from > > > > > Japanese reading. So we need to provide information on > reading/pronunciation. > > > > > > > > > > > > > > > The rest looks good to me. John, thank you very much! > > > > > > > > > > > > > > > Cheers, > > > > > Makoto > > > > > > > > -- > Takayuki Watanabe <nabe@lab.twcu.ac.jp> > Dept. Communication, Tokyo Woman's Christian University > Tel: +81-3-5382-6524, Fax +81-3-5382-4920 -- WATANABE Takayuki <nabe@lab.twcu.ac.jp> 東京女子大学現代文化学部コミュニケーション学科
Received on Monday, 6 March 2006 17:14:17 UTC