Talk:SC1-1-1-text-alternative

Open issues

Ready for review

Closed issues

• Typo in help text step 18: ‘anwer’.

Twan van Houtum (talk) 9:30, 22 April 2015 (UTC)

• Added new steps 4 to 6: check for ARIA10 (missed that).

Frank Berker (talk) 10:11, 9 April 2015 (UTC)

• Step 4: Should we handle decorative groups of images? Worlds worst web-designer may include a bunch of images as a background. - Should be covered by the new Step 4. If the group is not assessed as functional, all images are separated handled.

Frank Berker (talk) 10:47, 26 March 2015 (UTC)

• Should we include a test step for too long short text alternatives. What's the threshold? - Denied. Done

Frank Berker (talk) 10:47, 26 March 2015 (UTC)

• Step 4: Check for technique procedure step 2 and step 3 (G196)? (Step 2: Check that the other items in the group are marked in a way that can be ignored by assistive technologies.) - in some scenarios a concatenation of the alternatives may make sense. Redundant alternatives for all images should be identified by the users as non-sense. Agreed. Done

Twan van Houtum (talk) 13:50, 12 March 2015 (UTC)

• Typo in the fourth assumption: ‘heigth’.

Twan van Houtum (talk) 9:26, 16 April 2015 (UTC)

• Selector: Removed the [not(@longdesc)] restriction on images. This means that the selectors for the 1.1.1.-tests are not disjunct anymore, but I think images providing longdesc also need a short text alternative.

Frank Berker (talk) 10:47, 26 March 2015 (UTC)

• With "User expertise and skills" you suggest "basic understanding of HTML", let's discuss - Copy and Paste-Error. Changed and Done
• I think these questions should all have a help text - Done

Wilco Fiers (talk) 14:19, 19 March 2015 (UTC)

• Assumption: Add an assumption about grouped img elements - Done
• Selector: The text "Steps below must be performed for each element matching the selector" is a good one. But may give the impression that this is only true for this test case. We should either use this in every selector of every test case, or not at all and explain this elsewhere. - removed. Done
• Step 2: To avoid confusion, restructure the sentense as followed: "Check if one of the following attributes is present: aria-labelledby with valid destinations or an alt, aria-label or title. - Done
• Step 8: Add an assumption about this size - Done
• Step 9: "this information must also be written in the textual alternative" The question is not about the text alternative, so it should not be mentioned. Only suggest that if those situations are the case the image is not considered decorative - Done
• Step 9: Should we also test if the alternative is adjecent to the image? - added as Step 15
• Step 13: "is an image" change to "is an img" - Done
• Step 13: Combine this with step 14. I think it makes sense - Done
• Step 15: You are presenting two items. Should we instead include the alt attribute as some value in the test question? - introduced variable T1. Done
• What do we do about images with longdesc or images that require longdesc - separate test case. Edited selector. Done
• Activity diagram: Can you take out the "can't tell" results, as these are not returned at that point? - Done
• Step 9 / 12 / 15: Use "image" instead of "element" in the test question. - rejected, because area, object and embed are also running through. Done

Wilco Fiers (talk) 14:40, 12 March 2015 (UTC)

• Add a back to Success Criterion link on the test case page. - Done

Twan van Houtum (talk) 13:50, 12 March 2015 (UTC)

• Steps 8,9,10,13,14: I'm not sure about the outcome identifier.

8 and 9 result in passed3, - changed. Done

13 and 14 result in passed4, because they check the same technique or failure. - steps combined. Done

On the other hand, 10 results in fail5, whilst checking for 2 different failures. - agreed to leave it like it is. Done

• Step 13: At the moment all elements are checked for F38, which explicitely mentions images. - agreed to leave it like it is. Done

Frank Berker (talk) 12:21, 6 March 2015 (UTC)

• We should name the steps. I've given names to steps in a few test cases, I think it's worth doing so here. - Done
• Step 1: This structure confused me a little, please clarify - Added extra step for element selection - Done
• Step 2, change "Check if <img> elements have" to Check if the selected element is an img element and has. - Done
• Step 3: I'm not sure what the help text should be here. Please clarify - Agreed that this is the implementors decision. Added note to decision support page. - Done
• Step 3: I think the note should be changed and be it's own action in this step. - Done
• Step 6: You can fail automatically here if there is no other text in the link - Done
• Step 8: Link refers to step 6, but this should be step 9 - Done

Wilco Fiers (talk) 14:51, 5 March 2015 (UTC)

• incorporate F20: Failure of Success Criterion 1.1.1 and 4.1.2 due to not updating text alternatives when changes to non-text content occur - Decision: not checked (web page state) - Done

Frank Berker (talk) 11:50, 5 March 2015 (UTC)

• Step 7 (Version 0): This should probably mention links or other interactive components

Wilco Fiers (talk) 13:47, 29 January 2015 (UTC)

• Step 8: Link refers to step 6, but this should be step 9 - Done

Wilco Fiers (talk) 14:51, 5 March 2015 (UTC)

• derive definition/ hints concerning "sufficient description" from http://www.w3.org/WAI/tutorials/images/decision-tree/ - Done
• do the check for non-empty early in the TC (step 2?) - Done
• add step checking missing alt & not role="presentation" - Done
• add step checking no alternative for image contained in a link - Done
• add step checking for spacers - Done
• check processing (reachability) of images in groups - Done
• add the hint of step 7 to step 3 - Done
• add assumption: users vision needed - Done
• check possibility to split test in images / others - Done, would cause a repetition of test steps

Frank Berker (talk) 08:05, 30 January 2015 (UTC)

• Should we also look at links with images without a text alternative? - Yes, in that cases the alternative may be empty - Done
• Should the test fail if there is no alt attribute on img elements and there is no other alternative? - Yes, according to F65 - Done
• Test properties: The test mode should be SemiAuto - Done
• Test properties: We should limit the manual tests to sighted users - Done
• Step 3: This step looks at multiple images at once. We should ensure that this question should only be asked once for the entire group. - Done
• Step 3: Please clearify what "sufficiently discribes" means. - Done
• Step 4: Create an algorithm for "filename" - Done
• Step 6: What about decorative images here? - Done

Wilco Fiers (talk) 13:47, 29 January 2015 (UTC)

• Step 2: This step should also use the Text Alternative Computation Algorithm. The condition should be "Check if the <img> element has adjacent siblings of type <img>, of which one has a textual alternative (as computed by the Text Alternative Computation Algorithm." To catch the non-empty condition, add a step similar to Step 5 before Step 3. - Done
• Step 2: Clarification needed exactly one? - No! Done
• Step 3: The Text Alternative Computation Algorithm doesn't compute the alternative for a group of images. - Added to test step 3 - Done
• Step 7: Avoid negation in user input questions ("... does not contain information")? The explanation of decorative can be given in the help text. - Done

Annika Nietzio (talk) 14:30, 29 January 2015 (UTC)