- From: Gregg Vanderheiden <gv@trace.wisc.edu>
- Date: Tue, 28 Dec 2004 12:40:35 -0600
- To: "'John M Slatin'" <john_slatin@austin.utexas.edu>, <w3c-wai-gl@w3.org>
- Message-ID: <auto-000198992846@spamarrest.com>
Hi John, For ease - i will just put numbers below in brackets and put the answers up here. 1) Yes - this whole guideline (and all of the programmatically located provisions) is based on the cascading dictionary concept (or a comprehensive dictionary). When we put these in we said that they could stay only as long as we were able to solve that. but if we arent - then we need to remove them -- not change them to programmatically determined. That has not been discussed at all. 2) RE User agent issue. There is a user agent component to this in that the user agent would have to do something with the information provided. But user agents cannot distinguish between different definitions in different locations or provide definitions for domain specific words like "StickyKeys" etc. the Author would have to provide pointers to sources of definitions and pronunciations. 3) RE comment regarding pronunciation being in the SC. -- the question wasnt whether pronunciation was in the SC. You are correct - it was. but it was only programmatically located - not programmatically determined. that was the problem. 4) We could separate pronunciation and definition. That wouldnt make sense at Level 1 but it could at level 2 if we allow partial declarations which i think we will. they would mostly be solved with the same mechanism in most cases-- but not necessarily all. 5) RE 'Phrase" vs "run of text". The problem is that "run of text" includes both words and phrases. the SC only covers words. so the Phrases part of 'run of text' is the part that is beyond the SC. 6) you are correct about Japanese. They don't really have what we call 'words' though they do have characters and sentences. Hmmm. interesting problem. ALso in Dutch we have an ambiguity since they create words out of phrases. too much to think about between holidays !! John - you are a slave driver. Gregg -- ------------------------------ Gregg C Vanderheiden Ph.D. Professor - Ind. Engr. & BioMed Engr. Director - Trace R & D Center University of Wisconsin-Madison _____ From: John M Slatin [mailto:john_slatin@austin.utexas.edu] Sent: Tuesday, December 28, 2004 11:45 AM To: Gregg Vanderheiden; w3c-wai-gl@w3.org Subject: RE: [Techniques] Drft General Technique for GL 3.1 L2 SC1 Gregg wrote: <blockquote> I think the techniques here though differ from the SC that they are tied to. The proposed text is both more and less than the Success criterion cited requires. It is more in that the SC only requires that pronunciation be locatable. If there are multiple pronunciations - the guideline does not require you to say which one is appropriate (that would be 'programmatically determined'. So much of the text below does not apply. At least not for this SC. </blockquote> I'd be grateful to anyone who can give us a technique for programmatically locating pronunciation information for all words in the content! I suppose one technique could be to write a script that automatically submits a user-selected word to an online dictionary and goes directly to the section of the dictionary entry that describes pronunciation-- like the way the Valid HTML icon sends the page to the HTML Validation Service. [1] Some of the online dictionaries offer free toolbars that allow the user to key or paste in a word which is then automatically submitted to that particular dictionary site. Does the availability of such tools make this criterion unnecessary? (This goes back to a question Wendy raised several months ago shen she asked if this was a user agent issue and not a content issue.) [2] Gregg continues: <blockquote> The Level 3 success criteria for this guideline does specify that the meaning should be indicated but not the pronunciation. But that would not go here for this Level 2 SC and it is meaning - not pronunciation (though you might be able to work backward). [jms] </blockquote> I don't understand. Level 2 Success Criterion 1, as it appears in the 19 November public working draft, specifically talks about meanings and pronunciations: <current> The meanings and pronunciations of all words in the content can be programmatically located. </current> [3] I'm not certain, but it might be a good idea to separate meanings and pronunciations into separate success criteria (i.e., modify the guideline) so we can clarify exactly which techniques are appropriate/required for pronunciation and which are appropriate/required for meaning. [4] [jms] Gregg again: <blockquote> It is also more since it requires this for phrases. But the SC in consideration is only for words </blockquote> Sorry, Gregg, but I don't see where the draft below imposes requirements for "phrases." It does talk about a "run of text," a phrase I lifted directly from the Ruby Annotation specification (http://www.w3.org/tr/ruby/). A run of text may be a single word, or, in East Asian languages, a single character; it *could* be a phrase, I guess, but I was careful not to use that word in the draft technique. (I did use the word "phrase" in another draft technique about language-changes, so maybe that's what you're thinking of?) [5] [jms] Gregg again: <blockquote> It is less than guideline because it says "where meaning depends on pronunciation" Yet this SC does not restrict itself to any particular words or phrases. It requires that all words be locatable. </blockquote> You're right; thanks. [jms] It occurs to me that this SC, as written, may be problematic for languages like Japanese, where it may be necessary to provide information about how specific *characters* should be pronounced. So here again we may need to revise the success criterion to make room for languages that operate very differently from English (or other languages that use a Roman alphabet, etc.). [6] Gregg again: <blockquote> Remember - programmatically locatable just means that you get a list of meanings and the correct one is there somewhere. (like when you look a word up in a dictionary) You need 'programmatically determined' to have it tell you which meaning (or pronunciation) is the correct one of the bunch. </blockquote> I understand. Frankly I don't know how to make this sort of information programmatically locatable, and I put this out here in the hope that some brilliant person on the list would offer a solution (and write up the technique!) <grin> <blockquote> Thanks again [jms] </blockquote> You're welcome! Gregg [jms] John -- ------------------------------ Gregg C Vanderheiden Ph.D. Professor - Ind. Engr. & BioMed Engr. Director - Trace R & D Center University of Wisconsin-Madison _____ From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On Behalf Of John M Slatin Sent: Monday, December 27, 2004 5:40 PM To: w3c-wai-gl@w3.org Subject: [Techniques] Drft General Technique for GL 3.1 L2 SC1 The proposal below is part of the first draft of material for the General Techniques for Guideline 3.1, the guideline that contains some key requirements about language use. It's my hope that this material can be included in the next internal working draft, and that it will eventually make its way-- duly modified and corrected-- into the next public working draft. Guideline 3.1 L2 SC1: requirs: <current> The meanings and pronunciations of all words in the content can be programmatically located. </current> It would be very helpful if people with knowledge of writing systems for languages that do not use Roman or romanized alphabets would review and make suggestions for corrections, additions, deletions, etc. <proposed> Short-name for this technique: Pronunciation for users Task Information about the pronunciation of a run of text is explicitly associated with the run of text where meaning depends on pronunciation. Description There are many languages in which a run of text may mean different things depending on how the text is pronounced. This is common in East Asian languages as well as Hebrew, Arabic, and other languages; it also occurs in English and other Western European languages. Users with disabilities that make it difficult to use contextual cues as a guide to pronunciation and meaning benefit when information about how to pronounce potentially ambiguous text is available. Techniques for associating content with information about pronunciation vary depending upon the type and language of the content. For example, Ruby Annotation is appropriate for indicating pronunciation in some languages, such as Japanese, Chinese, and Korean. However, Ruby may be unnecessary in languages where Unicode fonts can include diacritical marks that convey pronunciation. Ruby Annotation allows the author to annotate a "base text," providing both a guide to pronunciation and, in some cases, a definition as well. Ruby is commonly used for text in Japanese and other East Asian languages. Ruby Annotation is defined as a module for XHTML 1.1. There are two types of Ruby markup: simple and complex. Simple Ruby markup applies to a run of text such as a complete word or phrase. This is known as the "base" text. The Ruby annotation that indicates how to pronounce the term is usually displayed immediately before the base text, and is shown in a smaller font. (The term "Ruby" is derived from a small font used for this purpose in printed texts.) Simple Ruby markup also provides a "fallback" option for user agents that do not support Ruby markup. Complex Ruby markup makes it possible to associate a single base text with more than one annotation. In such cases, the first annotation would typically indicate pronunciation and the second would provide the meaning. Complex Ruby markup also makes it possible to divide the baste text into smaller units, each of which may be associated with a separate Ruby annotation. Complex Ruby markup does not support the fallback option. Note: The primary reason for indicating pronunciation through Ruby or any other means is to make the content accessible to people with disabilities who can read and understand the language of the content if information about pronunciation is provided. Creating explicit association between the content and the pronunciation information ensures that pronunciation information remains available if the presentation format is adapted to meet the user's needs. Editor's note: Complex Ruby markup may be sufficient to satisfy this success criterion when pronunciation and meaning are provided in separate annotations of the same base text Editor's Note: As an additional benefit, it has also been suggested that Ruby Annotation might be used to make content accessible to people who use symbolic languages together with or as an alternative to conventional text. For example, a symmbol image could be used as a Ruby annotation above a base text. Such practices might benefit people whose speech or reading are impaired as the result of stroke or other injury to the brain, or from other causes. (See A. Judson, M. Lundalv, B. Farre, and L. Nordberg, <a href="http://dewey.computing.dundee.ac.uk/ccf/cop/#d0e876" <http://dewey.computing.dundee.ac.uk/ccf/cop/#d0e876> >Concept Coding Framework</a>) However, the Ruby 1.0 Specification does not support use of images, so implementation of this suggestion would depend upon a change in the Ruby specification. Resources <a href="http://www.w3.org/TR/ruby/" <http://www.w3.org/TR/ruby/> >Ruby Annotation</a> <a href="http://ncam.wgbh.org/salt/guidelines/sec11.html" <http://ncam.wgbh.org/salt/guidelines/sec11.html> >IMS Guidelines for Topic-Specific Accessibility</a> HTML Techniques <a href="http://www.w3.org/TR/WCAG20-HTML-TECHS/#lang-att_change" <http://www.w3.org/TR/WCAG20-HTML-TECHS/#lang-att_change> >Identifyin g language changes</a> CSS Techniques <a href="http://www.w3.org/TR/css3-ruby" <http://www.w3.org/TR/css3-ruby> >CSS 3 Ruby</a> </proposed> "Good design is accessible design." Dr. John M. Slatin, Director Accessibility Institute University of Texas at Austin FAC 248C 1 University Station G9600 Austin, TX 78712 ph 512-495-4288, fax 512-495-4524 email jslatin@mail.utexas.edu Web <http://www.ital.utexas.edu/> http://www.utexas.edu <http://www.utexas.edu/research/accessibility> /research/accessibility
Received on Tuesday, 28 December 2004 18:41:02 UTC