Minutes for Teleconference on 9 October 2014

Automated WCAG Monitoring Community Group
Teleconference 2014-10-9, 16:00 - 17:00 h (CET)

Participants: Wilco, Birkir, Hanno, Kamyar, John, Annika, Katie, Frank 
(scribe)

ITEM 1 Action items: 
https://www.w3.org/community/auto-wcag/wiki/Action_items
Annika: An explanation of language identification algorithms [1] is 
added to SC3-1-1-text [2] . None of the available algorithms seem to be 
sufficient. Hanno suggests a benchmarking of algorithms. Annika will 
update the Wiki-page with some suggestions and will add some assumptions.
Birkir: Tested behaviour of screen-readers concerning the 
xml:lang-attribute vs. the lang-attribute.  The latter is used in all 
cases. VoiceOver-tests are pending.
Wilco: Added a definition of non-empty [3] to SC4-1-2-button [4].


ITEM 2+3: Finalizing 3.1.1 Language of Page [5] and 3.1.2 Language of 
Parts [6]
Annika: Changes in the lang attribute selector are needed.
Annika, Katie and Birkir discuss  about the lang attribute and its usage 
by assistive technologies. Screen readers use the lang attribute 
independent of the real language of the content. Strange behaviour on 
the title attribute, which is always read in the language set in the 
browser/system.
Annika asks, if the test of both attributes, xml:lang and lang, equality 
could be skipped. Katie and Birkir advise against it.
Decision: Keep the step.

ITEM 4: Review 2.4.2 Page Titled [7]
1. Step 3: "Check if the title is a filename or a URL.":
Wilco: How should a text be checked to be not a filename?
Annika: Titles can be filenames, e.g. on Google Docs.
Hanno: A lookup for a file extension can be done.

2. General:
Wilco: The seclector should better be <html> than <head>, because the 
test object is the entire page and the selectors are for the test of 
individual elements.
Annika: The Selectors are mainly meant for the programmers, so using the 
<head> selector is correct.
Decision: Keep <head> as selector.

Wilco: Test for non-empty-text is missing.

3. General discussion on range of single tests
Wilco: Tests should maybe be splitted into automatic and manual tests.
Birkir: The tests are items of the same testing process. Firstly, the 
automated tests are done. Afterwards, things that could not have been 
decided automatically, are tested manually.
Wilco: O.K., but we should try to keep the tests small, but logically 
connected.

4. General discussion on tests on mobile devices
Katie: Are we checking also on mobiles? There is no tab on mobiles and 
therefore no title displayed.
Annika: Also mobile devices use the title.
Birkir: Agrees. The title is read by screen readers. Suggestion: On 
mobile devices, display the title explicitely to the users in the 
testing tool.

ITEM 5: Brainstorm on auto-wcag communication to the interested public
Birkir: Submitted a presentation proposal for CSUN.
Birkir: Suggests use of social-media, e.g. Twitter

ITEM 6: Final comments
Hanno: Question about feedback on SC1-2-2-video-element-without-track 
and SC1-2-2-video-element: Splitting the tests is quite difficult. Will 
send a separate mail.


REFERENCES:

[1] 
http://www.w3.org/community/auto-wcag/wiki/Language_identification_algorithm
[2] https://www.w3.org/community/auto-wcag/wiki/SC3-1-1-text
[3] https://www.w3.org/community/auto-wcag/wiki/Non-empty
[4] https://www.w3.org/community/auto-wcag/wiki/SC4-1-2-button
[5] https://www.w3.org/community/auto-wcag/wiki/3.1.1_Language_of_Page
[6] https://www.w3.org/community/auto-wcag/wiki/3.1.2_Language_of_Parts
[7] https://www.w3.org/community/auto-wcag/wiki/2.4.2_Page_Titled



SUMMARY OF ACTIONS:

https://www.w3.org/community/auto-wcag/wiki/index.php?title=Action_items&oldid=708

Received on Saturday, 11 October 2014 07:27:30 UTC