See also: IRC log
http://lists.w3.org/Archives/Public/w3c-wai-au/2012AprJun/0060.html
http://www.w3.org/WAI/AU/2012/ATAG2-CommentResponses20124010LC-3.html
http://www.w3.org/WAI/AU/2012/ATAG2-CommentResponses20124010LC-3.html
AUWG: The Working Group agrees that this could be clarified and that the stem text should change to better synch with B.2.3.3 (the style also follows that employed for the stems of B.1.1.1 and B.1.1.2).
The text has been reworded as follows: "Repair of Text Alternatives During Authoring Sessions: If the authoring tool attempts to automatically or semi-automatically repair text alternatives for non-text content ("repair strings") during an authoring session, then the following are both true: (Level A)
(a) Suitable Text Sources: Repair strings are only ever derived from text sources designed to fulfill the same purpose as the text alternative (e.g., suggesting an image's "description" metadata field as a long description). Other text attributes (e.g., the file name, file format) or generic strings (e.g. "image") are not used.
(b) Author Control: Authors have the opportunity to accept, modify, or reject the repair strings prior to insertion in the content; and
JR: No change in meaning, just removing connotation and following pattern of other stems
AL: Sounds good
AUWG: The Working Group agrees that this could be clarified and that the stem text should change to better synch with B.2.3.3 (the style also follows that employed for the stems of B.1.1.1 and B.1.1.2).
The text has been reworded as follows: "Repair of Text Alternatives During Authoring Sessions: If the authoring tool attempts to automatically or semi-automatically repair text alternatives for non-text content ("repair strings") during an authoring session, then the following are both true: (Level A)
(a) Suitable Text Sources: Repair strings are only ever derived from text sources designed to fulfill the same purpose as the text alternative (e.g., suggesting an image's "description" metadata field as a long description). Other text attributes (e.g., the file name, file format) or generic strings (e.g. "image") are not used; and
(b) Author Control: Authors have the opportunity to accept, modify, or reject the repair strings prior to insertion in the content.
Resolved: All agree on language immediately above.
AUWG: "Relevant Sources" is only
the handle, not the testable part of the condition. See
rewording under MS3.
... "Relevant Sources" is only the handle, not the testable
part of the condition. See rewording under MS3.
Resolved: All agree on the response immediately above.
<Zakim> jeanne, you wanted to say that since pattern and context recognition is continually improving, I would not want to see us forbid it.
<jeanne> Jan: But then a checker will give a pass to the author, even though they have not created meaningful alt text.
<jeanne> ... but examples such as Facebook automatically flagging a photo as "Jan's Profile Picture"
<jeanne> ... but it would still fail the part where the user has to approve it.
<jeanne> ... perhaps it could be flagged as automatically repaired, and that could be brought to the author's attention when the author ran the checker.
<jeanne> ... it could be a note.
<jeanne> +1 for the note
Possible note: If the tool does auto-correct during the session, the alt needs to be flagged as auto-generated...
<jeanne> Greg: the author would need to have their attention drawn to a change in content, to make sure the system did it correctly.
<jeanne> ... and upon initial content creation -- an opportunity to verify that it is, in fact, meaningful.
<jeanne> Jan: B.2.3.2 - a type of repair that is "almost" suitable text source, but it needs to be verified by the author.
<scribe> ACTION: JR to write note re: If the tool does auto-correct during the session, the alt needs to be flagged as auto-generated. [recorded in http://www.w3.org/2012/06/25-au-minutes.html#action01]
<trackbot> Created ACTION-381 - Write note re: If the tool does auto-correct during the session, the alt needs to be flagged as auto-generated. [on Jan Richards - due 2012-07-02].
"Repair of Text Alternatives During Authoring Sessions: If the authoring tool attempts to automatically or semi-automatically repair text alternatives for non-text content ("repair strings") during an authoring session, then the following are both true: (Level A)
(a) Suitable Text Sources: Repair strings are only ever derived from text sources designed to fulfill the same purpose as the text alternative (e.g., suggesting an image's "description" metadata field as a long description). Other text attributes (e.g., the file name, file format) or generic strings (e.g. "image") are not used; and
(b) Author Control: Authors have the opportunity to accept, modify, or reject the repair strings prior to insertion in the content.
Resolved: Alll accept the text above for B.2.3.2.
B.2.3.3 Repair of Text Alternatives After Authoring Sessions: If the authoring tool attempts to automatically repair text alternatives for non-text content after an authoring session has ended, then any repair strings do not include text that is also available to user agents (e.g. the file name, metadata stored in non-text content) or generic text strings (e.g., "image").
Note: Examples of acceptable repair strings include those derived from contextual information (e.g., that an image is the author's profile picture) or from performing pattern recognition on the non-text content. (Level A)
JT: Does is this address awkwardness of wording?
B.2.3.3 Repair of Text Alternatives After Authoring Sessions: If the authoring tool attempts to automatically repair text alternatives for non-text content after an authoring session has ended, then must do so from text sources that are otherwise still available to user agents (e.g. the file name, metadata stored in non-text content) or generic text strings (e.g., "image").
B.2.3.3 Repair of Text Alternatives After Authoring Sessions: If the authoring tool attempts to automatically repair text alternatives for non-text content after an authoring session has ended, then it must do so only from text sources that are otherwise still available to user agents (e.g. the file name, metadata stored in non-text content) or generic text strings (e.g., "image").
B.2.3.3 Repair of Text Alternatives After Authoring Sessions: If the authoring tool attempts to automatically repair text alternatives for non-text content after an authoring session has ended, then it must not do so from text sources that are otherwise still available to user agents (e.g. the file name, metadata stored in non-text content) or generic text strings (e.g., "image").
AL: The text is an improvement...I'd like tot think about it more.
<jeanne> jeanne: wants to add an additional example (e.g. the file name, metadata stored in non-text content), generic text strings (e.g., "image") or marking the image as decorative.
JT: Let's cycle back to this next week
AUWG: While the Working Group recognizes that "range" is a weak term, the Working Group considered stronger language, but no workable solution was found. For example, if we required an accessible template for every use case, would we be prepared to fail an entire tool if it included many accessible templates, but lacked one for a challenging domain, such as a calendar? The Working Group...
<jeanne> +1 to the response
AUWG: decided to use wording that clearly conveyed our intent while remaining testable.
JT: Could we move to addressing this, by putting our strong support in as a note or other measure
<jeanne> +1
AUWG: While the Working Group
recognizes that "range" is a weak term, the Working Group
considered stronger language, but no workable solution was
found. For example, if we required an accessible template for
every use case, would we be prepared to fail an entire tool if
it included many accessible templates, but lacked one for a
challenging domain, such as a calendar? The Working
Group...
... decided to use wording that clearly conveyed our intent
while remaining testable.
We have repeated the informative note under "Range" ("Note: ATAG 2.0 uses the term "range" in several success criteria in which absolute measurements may not be practical (e.g., the set of all help documentation examples, the set of all templates). While the strict testable requirement is the definition "More than one item within a multi-item set", implementers are strongly encouraged to...
scribe: implement the success criteria more broadly.") under the intent sections of all of the SCs where it is used.
Resolved: Accept text above (response and changes to the various intent sections)
AUWG: A note will be added:
Note: The nature of the accessibility status indicator is not specified and will depend on the type of content. For example, a widget gallery might indicate a WCAG 2.0 conformance level for each widget, while a clip-art gallery might simply indicate whether an alt text string is provided for each image.
SN: Sounds fine.
Resolved: All accept "Note: The nature of the accessibility status indicator is not specified and will depend on the type of content. For example, a widget gallery might indicate a WCAG 2.0 conformance level for each widget, while a clip-art gallery might simply indicate whether an alt text string is provided for each image."
<jeanne> http://www.w3.org/WAI/AU/2012/ATAG2-CommentResponses20124010LC-3.html
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) No ScribeNick specified. Guessing ScribeNick: Jan Inferring Scribes: Jan Default Present: Jeanne, Jan, Alex, Cherie, Jutta, +1.571.765.aaaa, Greg, +1.561.582.aabb Present: Jeanne Jan Alex Cherie Jutta +1.571.765.aaaa Greg +1.561.582.aabb Regrets: Tim B. Agenda: http://lists.w3.org/Archives/Public/w3c-wai-au/2012AprJun/0060.html Got date from IRC log name: 25 Jun 2012 Guessing minutes URL: http://www.w3.org/2012/06/25-au-minutes.html People with action items: jr[End of scribe.perl diagnostic output]