October 28, 1999
4:00 - 5:00 PM EDT
Longfellow Bridge: (617-252-1038)
ian jacobs, wendy chisholm, charles mccathieneville, jason white, chuck letourneau, frank torrey, tim curtin, al gilman
regrets: gregory rosmaita, gregg vanderheiden, and william loughburough
a. Is blank alt text " " forbidden always? If so Why? b. Is blank alt text " " forbidden for purposes of spacing text apart? If so, why? c. Is " " forbidden always? If so, Why? d. Is " " forbidden for purposes of spacing text apart? If so, Why?
/* introductions */
JW &nsp; is not allowed in cdata since entities are not allowed in cdata.
IJ according to spec you can have entities in cdata. "authors should not declare...browser can ignore."
JW therefore, UA can ignore.
CMN which reinforces the argument that spaces in alt for presentation effect is not right.
IJ nothing to do with image, may be ignored.
WC won't hurt anything, so can't use?
CMN but will if not rendered.
JW certain situations, like braille or speech device, multiple spaces would be ignored anyway.
CMN reads from techniques doc - example. should not be deprecated, should be taken out. it's a bad example.
Resolved: shouldn't use " ".   just another form of white space. " " because ignored.
IJ is a useful thing to do when you don't want a break, but should not be used for formatting.
WC e.g. WCAG keep priority and number together.
FT can use in tables.
IJ can't rely on it. just so happens that in neetscape you'll get a border.
WC argument for not using in alt.
JW using it to insert a space, should be outside of attribute. wnat something more directly to image.
AG if the image is a space, then may get too much space in viz presentation. i thought only use of was when image was a spacer.
WC but null alt-text.
AG unless holding space between words.
CL but use alt-space.
AG that's semantics of using "deprecated" example.
TC running words together or repeating them is a problem.
WC ER's purposes.
CL also because example is very confusing.
/* discussing <A href="foo.html"> <img alt="" src=...> prices </A> */
CL or alt="link to"
resolved: o.k.
/* <A href="foo.html"> <img alt="prices" ...><img alt="" src=...> </A> - weird. abuse of image. */
FT why not leave alt off?
group - because required, helps automatic validators.
resolved: not o.k. abuse of IMG.
CL another concern w/5.6.1 alt="spacer"
FT good application for null alt-text.
@@editors need to fix examples of the use of null alt-text and spaces in alt-text.
JW delete last one. replace earlier with alt=" " not allowed.
CL can go into errata.
JW nope, just for guidelines.
Tim alt only used if content?
CMN no, only when meaningful. if not meaningful, use null.
CL still people that would disagree. but we seem to agree. meaningless graphics should not have meaning.
Tim we still have issues with tables. where they have multi-cols and rows, same data in each cell. if user navigates from cell to cell they can become lost. thoughts or discussion on tag that might equate to col heading or row heading.
JW yes, HTML 4.0 defines several attributes. can make explicit connection.
/* Tim also wants to know what he should tell I.R.S. employees who are blind */
/* discussion of user agents, html attributes, plug-ins and other repair tools such as table linearizer */
Tim what if collaborative effort between IE and Jaws so that certain key sequence will grab heading and cell.
JW UA guidelines suggest how to do.
WC trace tools do this.
AG perhaps should generate alternate page.
TC with all of the technology we could implement tools so we don't have to do special things.
WC ER!
AG lots of brains not all connected with your customers.
CMN serifs.
CL never seen.
JW when I18N have been worked out, but what already be deprecated be even more so. the intention was to allow an AT to distinguish links.
CMN already things that read non-roman alphabets.
AG but they know what they are reading. space useful in roman languages.
JW what about greek or russian?
CMN same typography.
CL cop-out, but "for languages that use spaces, use spaces."
IJ this is a UUA? can we get rid of?
CMN pure screen reader that doesn't associate any text. jason - used lynx on a brailler? how does it do?
JW yes. a case where no space is rare. i'd have to look at it.
@@CMN put up a test case, have savvy and non-savvy users look at.
JW home page reader would do it correctly.
@@CMN find out which user agents do not separate links with a space or which ATs read links together.
AG yes, or chuck's fix.
IJ space as input, space as ouput. here space as output. space as input may be required for parsing.
/* people don't understand comment */
AG translates have we got browsers to implement.
CMN yes but won't save in practice.
unresolved: until get results from CMN action items.
CMN wrote DD
WC so include in errata so people can use?
CMN sent to GL a week after REC.
JW is it acceptable to use. should it be used or permitted be used as substitute for method of conformance claim outlined in document.
CMN agree way to claim conformance. do people feel a sufficient way to claim?
WC what are arguments against?
CMN not in the Rec.
WC b/c harder to determine if people did claim?
JW if UA doesn't support PICS or RDF.
AG closet conformant?
WC a new technology, therefore needs to transform gracefully?
CMN not new.
AG not used.
CMN yes, in netscape.effectively: warn me when i access something that exceeds rating.
AG implemented in search engines?
CMN yes. have a look at w3.org/pics which describes where and how.
CL we don't have a checkpoint that says, must indicate conformance.
IJ if do in closet, can't market. what's the impact?
JW statement that conformance claims made in particular manner. CMN is saying that meta data be constitute another way to claim and be sufficient.
CL agrees.
JW problem b/c text-based UAs will have difficult time recognizing.
CMN hairy issue: what UAs do we mean by UA?
IJ you mean in "UUA?"
CMN yes.
IJ of the 12, 4 or 5 for ATs.
resolved: finish discussing use of meta data to claim conformance at next meeting.
resolved: postpone next meeting.
CMN do we meet on U.S. Thanksgiving?
/* resolved: those that wants to meet on thanksgiving will meet */
/* resolved: we will also meet on veterans day */
Copyright © 1999 W3C (MIT, INRIA, Keio), All Rights Reserved. W3C liability, trademark, document use and software licensing rules apply.