- From: Chris Ridpath <chris.ridpath@utoronto.ca>
- Date: Thu, 5 Aug 1999 16:27:10 -0400
- To: <w3c-wai-er-ig@w3.org>, "Leonard R. Kasday" <kasday@acm.org>
Len, Thanks for your helpful comments. I've put most of them into the document and will have the new version posted at the A-Prompt site tomorrow Fri. Aug. 6 (http://aprompt.snow.utoronto.ca/docs/Implementation.html). I have a few comments on your suggestions: > > It also describes methods for modifying HTML > >documents so that they will conform to these guidelines. > > LRK:: minor: for the software to modify... > CR:: I don't understand this. Are you suggesting "It also describes methods for the software to modify documents so that they will conform to these guidelines."? > LRK:: important: in all cases, even if ALT text passes tests, show picture > and ALT text so that user can verify that ALT text is suitable. > CR:: We are designing a 'review' or 'verify' mode so that the user can review/verify what is actually in the document. I think this is what you're suggesting. We really don't want to present the user with items that are OK when they are in 'evaluate/repair' mode and only questionable items should be shown. When we get the 'review' mode working it will address some of your following comments. > >Suggestions for possible ALT text: > > LRK:: minor: reword: "Ways that program can automatically suggest alt text" > CR:: Daniel has made a suggestion that we use the more generic 'repair technique' text for these sort of suggestions. OK? > > * IMG element should have a valid LONGDESC attribute if the image is > > complex. > > * If IMG element has no LONGDESC attribute and could be a complex image, > > ask user if the image is complex and requires a long description. > > LRK:: signficant: I suggest that this be folded in with display of images > and ALT text. Display description pointed to by LONGDESC, or alert user at > that time if no LONGDESC. This eliminates need to evaluate what's a > "complex" image or the bullet and rule line below. > CR:: For the 'review/verify' mode this makes sense. For 'evaluate/repair' mode I think we'll have to ask the user for every image (that's not a bullet or HR) whether it's complex and requires a LONGDESC. > >Technique 1.1.K [priority 3] User notification for ASCII art > > > LRK:: signficant: algorithm for spotting ASCII art, e.g. series of n or > more punctuation characters or clusters of repeated letters... > CR:: Developing the algorithm will take a lot of work. Anybody got some free time? > >Checkpoint 1.2 - Provide redundant text links for each active region of a > >server-side image map. > > > >Technique 1.2.A [priority 1] Prompt user for text links if ISMAP used. > > > > * If an IMG element has a valid ISMAP attribute, prompt the user for > > associated text links. > > LRK:: important: let user select purported equivalent links and check > against server links. The server side links may be obtained by having user > click on each apparently selectable region; or generated automatically by > sending grid of selection points to server (as done by some robots). > CR:: Good idea. I've heard that some servers can provide a set of links and associated coordinates for the image map if you ask for it. > >Checkpoint 1.3 - Until user agents can automatically read aloud the text > >equivalent of a visual track, provide an auditory description of the > >important information of the visual track of a multimedia presentation. > > > >(Can't be machine checked. User notification if movie file found?) > > LRK:: important: allow user to run track and listen to description. This > also applies to checkpoint 1.5. > CR:: This is a good idea. We would like to add a media player to A-Prompt for doing just this sort of thing. It likely won't get into this version though. Chris ----- Original Message ----- From: Leonard R. Kasday <kasday@acm.org> To: <w3c-wai-er-ig@w3.org> Sent: Monday, July 26, 1999 4:23 PM Subject: Techniques document: more comments on checkpoints 1.x > Here's some comments on the techniques document that chris is editing. I've > labeled comments as "minor: signficicant", "important" > > This has most text of original. I've preceded my comments with LRK:: so > you can quickly skip to them. > > This is for checkpoints 1.x, i.e. 1.1, 1.2 ... > > > ----------- > > > > It also describes methods for modifying HTML > >documents so that they will conform to these guidelines. > > LRK:: minor: for the software to modify... > > > >creation of software programs (A-Prompt, Bobby) and we want the software to > > LRK:: minor: add "e.g." before A-prompt... > > > >Guideline 1. Provide equivalent alternatives to auditory and visual > >content. > > > >Checkpoint 1.1 - Provide a text equivalent for every non-text element. > > > >Technique 1.1.A [priority 1] Check Images for ALT text > > > > * IMG element must have a valid ALT attribute. > > > >Valid ALT text: > > > > * Not allowed - NULL ALT value (ALT="") > > * Allowed - ALT value of 1 or more spaces (ALT=" ") > > LRK:: important: we're awaiting word from GL on null and blank alt text. > See discussion at > > http://lists.w3.org/Archives/Public/w3c-wai-er-ig/1999Jun/0050.html > > especially part about null or blank alt text for links. Please fold in > "suggestions" we made to GL pending their approval. > > > > * Suspicious - ALT attribute value could be file size (ends with > > "bytes") > > * Suspicious - ALT attribute value ends with image file suffix. > > * Suspicious - ALT attribute value is placeholder ALT text. > > LRK:: important: define placeholder, i.e. text supplied by generally used > or available HTML editors. > > LRK:: important: in all cases, even if ALT text passes tests, show picture > and ALT text so that user can verify that ALT text is suitable. > > >Language for missing ALT text: Missing ALT text for image > >Language for suspicious ALT text: Suspicious ALT text for image > > LRK:: significant: label as "suggested" language. We don't want to limit > designers here. For example, they may want to use icons (with appropriate > ALT text of course) > > > > >Suggestions for possible ALT text: > > LRK:: minor: reword: "Ways that program can automatically suggest alt text" > > > 1. If the document contains another instance of the image and that image > > contains ALT text, suggest that ALT text. > > LRK:: significant: if the _site_ contains another instance... > > > 2. If the width of the image and the height of the image are both less > > than 20 pixels, suggested text should be "bullet". > > 3. If the height of the image is less than 20 pixels and the width of the > > image is greater than 150 pixels, suggested text should be "horizontal > > rule". > > 4. Other suggestions by Daniel Dardailler > > 5. Suggestions by Michael Vorburger > > LRK:: signficant: give url's of Daniels and Michael's documents or better > yet, add to list.. > > >Other checks: > > > 1. After user has entered ALT text for the image, check the document for > > LRK:: signficant: check the _site_ , not just the document ... > > > other instances of the image. If the document contains other images > > that are the same and they do not have ALT text, suggest that all same > > images without ALT text use the new ALT text. > > > >Link to test files for this technique. > > > >Technique 1.1.B [priority 1] Check images for LONGDESC > > > > * IMG element should have a valid LONGDESC attribute if the image is > > complex. > > * If IMG element has no LONGDESC attribute and could be a complex image, > > ask user if the image is complex and requires a long description. > > LRK:: signficant: I suggest that this be folded in with display of images > and ALT text. Display description pointed to by LONGDESC, or alert user at > that time if no LONGDESC. This eliminates need to evaluate what's a > "complex" image or the bullet and rule line below. > > >Images that do not require a LONGDESC: > > > > * bullets - width and height of image are both less than 20 pixels > > * horizontal rules - height of image is less than 20 pixels and width of > > image is greater than 100 pixels > > > >Valid LONGDESC URI: > > > > * Any valid URI > > > >Language for missing LONGDESC: Complex images require a long description. > > LRK:: significant: this is global comment. In general, label all > "language" as "suggested language". > > > > >Link to discussion on this technique. > >Link to test files for this technique. > > > >Technique 1.1.C [priority 1] Check input buttons that use an image for ALT > >text > > > > * If an INPUTelement has a TYPE attribute with a value of "image" then > > the INPUT element must also have a valid ALT attribute. > > > >Valid ALT text: > > > > * Not allowed - NULL ALT value (ALT="") > > * Allowed - ALT value of 1 or more spaces (ALT=" ") > > > LRK:: important: ALT=" " not allowed for button. It's important to give > function of button. > > > * Suspicious - ALT attribute value could be file size (ends with > > "bytes") > > * Suspicious - ALT attribute value ends with image file suffix. > > * Suspicious - ALT attribute value is placeholder ALT text. > > LRK:: significant: see above comment for placeholdere ALT text. > LRK:: minor: instead of repeating these rules, put all in one place and > refer to them to avoid possible errors in future. Except of course for the > case of ALT="" or ALT=" " which is not allowed for buttons but is for > non-link text. > > > > >Language for missing ALT text: Missing image for this input element. > > > >Language for suspicious ALT text: Suspicious ALT text: (current ALT text) > > > >Link to test files for this technique. > > > >Technique 1.1.D [priority 1] Check applets for ALT text > > > > * APPLET element must have a valid ALT attribute. > > > >Valid ALT attribute text: > > > > * Not allowed - NULL ALT value (ALT="") > > * Allowed - ALT value of 1 or more spaces (ALT=" ") > > LRK:: important: again, see above discussion of null and blank alt text. > > > * Suspicious - ALT attribute value could be file size (ends with > > "bytes") > > * Suspicious - ALT attribute value ends with image file suffix. > > * Suspicious - ALT attribute value is placeholder ALT text. > > > >Language for missing ALT text: Missing ALT text for applet. > > > >Language for suspicious ALT text: Suspicious ALT text: (current ALT text) > > > >Link to test files for this technique. > > > >Technique 1.1.E [priority 1] Check Applets for alternative content > > > > * Between the APPLET start element and APPLET end element must be a > > valid text element. > > > >Valid text element: > > > > * Must contain at least one word of text > > * Suspicious - ALT attribute value is placeholder ALT text. > > > >Language for missing alternative content: Missing alternative content for > >applet. > > > >Language for suspicious alternative content: Suspicious alternative > >content: (suspicious text) > > > >Link to test files for this technique. > > > LRK:: important: provide quick means for user to view applet and > alternative content, preferably side by side, to check for equivalence. > This also applies to OBJECT. > > LRK:: important: provide means to check if applet itself is accesible. Do > we need to check both Microsoft and Sun/IBM methods of access? > > > >Technique 1.1.F [priority 1] Check objects for alternative representation > > > > * Between the OBJECT start element and OBJECT end element must be a > > valid text element. > > > LRK:: important. No. Alternative can be another object. For example, an > image with ALT text. > > > >Valid text element: > > > > * Must contain at least one word of text. > > * Suspicious - ALT attribute value is placeholder OBJECT ALT text > > > >Language for missing alternative representation: Missing alternative > >representation for this object. > > > >Language for suspicious alternative representation: Suspicious alternative > >representation: "suspicious text" > > > >Link to test files for this technique. > > > >Technique 1.1.G [priority 1] Check frames for an associated LONGDESC file > > LRK:: significant: see LONGDESC comments made above. > > > > * FRAME elements should have a valid LONGDESC attribute if the frame > > title does not completely describe the frame content > > * If FRAME does not have a LONGDESC attribute, ask user if frame > > contents are complex and requires one. > > > >Valid LONGDESC file name: > > > > * Must not be NULL > > * Must be a valid URI > > > >Language for missing LONGDESC: Missing 'long description' file for this > >frame. > > > >Language for suspicious LONGDESC: Suspicious 'long description' file name: > >"file name" > > > >Link to discussion on this technique. > >Link to test files for this technique. > > > >Technique 1.1.H [priority 1] Check AREA elements for ALT text > > > > * AREA elements must have a valid ALT attribute. > > > >Valid ALT attribute: > > > > * Not allowed - NULL ALT value (ALT="") > > * Suspicious - ALT attribute value is placeholder ALT text. > > LRK:: important: blank value ALT=" " not allowed either. > LRK:: signficant: carry over rest of ALT text discussion or better yet > point to it. > > > >Language for missing ALT text: Missing ALT text for this image map area. > > > >Language for suspicious LONGDESC: Suspicious ALT text for this image map > >area: (ALT text). > > > >Link to test files for this technique. > > > >Technique 1.1.I [priority 1] Check if audio files have an associated text > >transcript > > > > * If the target of an A element is a sound file then ask the user if > > there is an existing text transcript file. > > > >Language for suspected missing text file: Audio files require an associated > >text transcript file. Is there an associated text file for this audio file > >(audio file name)? > > > >Link to test files for this technique. > > > >Technique 1.1.J [priority 1] Check SCRIPT element for associated NOSCRIPT > >element > > > > * There must be a NOSCRIPT element immediately following a SCRIPT end > > element. > > * The NOSCRIPT start and end elements must contain at least one valid > > text element. > > > >Valid text element: > > > > * Must contain at least one word of text > > * Suspicious - ALT attribute value is placeholder NOSCRIPT text. > > > >Language for missing NOSCRIPT: Missing NOSCRIPT element for this SCRIPT > >element. > > > >Link to test files for this technique. > > > LRK:: important: allow user to compare script and noscript alternatives, > preferably side by side. > > > >Technique 1.1.K [priority 3] User notification for ASCII art > > > >This technique describes notifications presented to the user for > >checkpoints that can not be machine tested. > > > >BODY elements will generate the following user notification: If there is > >any ASCII art in this document, please give it a textual description. > > > >Link to discussion on this technique. > >Link to test files for this technique. > > LRK:: signficant: algorithm for spotting ASCII art, e.g. series of n or > more punctuation characters or clusters of repeated letters... > > > ------------------------------------------------------------------------ > > > >Checkpoint 1.2 - Provide redundant text links for each active region of a > >server-side image map. > > > >Technique 1.2.A [priority 1] Prompt user for text links if ISMAP used. > > > > * If an IMG element has a valid ISMAP attribute, prompt the user for > > associated text links. > > LRK:: important: let user select purported equivalent links and check > against server links. The server side links may be obtained by having user > click on each apparently selectable region; or generated automatically by > sending grid of selection points to server (as done by some robots). > > > >Valid ISMAP attribute: > > > > * Must be a valid URL > > > >Language for prompt: Server-side image maps should have associated text > >links. Are there text links in your document for this image map (image map > >file name)? > > > >Link to test files for this technique. > > > ------------------------------------------------------------------------ > > > >Checkpoint 1.3 - Until user agents can automatically read aloud the text > >equivalent of a visual track, provide an auditory description of the > >important information of the visual track of a multimedia presentation. > > > >(Can't be machine checked. User notification if movie file found?) > > LRK:: important: allow user to run track and listen to description. This > also applies to checkpoint 1.5. > > ------------------------------------------------------------------------ > > > >Checkpoint 1.4 - For any time-based multimedia presentation (e.g., a movie > >or animation), synchronize equivalent alternatives (e.g., captions or > >auditory descriptions of the visual track) with the presentation. > > > >(Can't be machine checked. User notification?) > > > ------------------------------------------------------------------------ > > > >Checkpoint 1.5 - Until user agents render text equivalents for client-side > >image map links, provide redundant text links for each active region of a > >client-side image map > > > >Technique 1.5.A [priority 3] Prompt user for text links if USEMAP used. > > > > * If an IMG element has a valid USEMAP attribute, prompt the user for > > associated text links if the document does not already contain > > associated text links. > > * Associated text links may be found by searching the document for > > anchors with HREF values that correspond to the AREA elements in the > > given USEMAP. > > > >Valid USEMAP attribute: > > > > * Must be a valid URL > > > >Language for prompt: Client-side image maps should have associated text > >links. > > > >Link to test files for this technique. > > LRK:: important: have use specify purported text links. Program then > checks against links in the image maps > > > > ------- > Leonard R. Kasday, Ph.D. > Universal Design Engineer, Institute on Disabilities/UAP, and > Adjunct Professor, Electrical Engineering > Temple University > > Ritter Hall Annex, Room 423, Philadelphia, PA 19122 > kasday@acm.org > (215} 204-2247 (voice) > (800) 750-7428 (TTY)
Received on Thursday, 5 August 1999 16:28:41 UTC