Talk:SC2-4-2-titles-across-pages

Open issues

• We should update the selector according to the multi-page setup we created
• Would it be possible for us to say that each page that doesn't include a form and that has different headings, it should have a unique title?

Wilco Fiers (talk) 13:52, 7 April 2016 (UTC)

• Step 2: Typo in first line: ntested.

Frank Berker (talk) 10:04, 7 April 2016 (UTC)

• Test properties table: the Success Criterion is missing. Change ‘Test requirement’ to ‘Success Criterion’ and add the link to the Success Criterion.

Twan van Houtum (talk) 12:08, 6 August 2015 (UTC)

• Test properties table: test name is missing
• The the value of TestCase on outcome tables needs be changed on steps 1,2,3. It should be SC2-4-2-titles-across-pages instead of 242-title-across-page
• The same for identifiers: change SC242-html-across-pages-fail1 to SC2-4-2-titles-across-pages-failed1. SC242-html-across-pages-fail2 to SC2-4-2-titles-across-pages-failed2. SC242-html-across-pages-fail3 to SC2-4-2-titles-across-pages-failed3 and SC242-html-across-pages-pass1 to SC2-4-2-titles-across-pages-passed1
• Step 1: after outcome table add: "else, continue with #Step 2"
• Step 2: add "if yes continue with #Step 3"

--Kamyar Rasta (talk) 14:38, 16 June 2015 (UTC)

• The description talks about "comparing titles of webpages within the same domain". There are other terms as well: "web site" or "set of Web pages". The latter is defined in WCAG 2.0. Which term are we using in auto-wcag.

--Annika Nietzio (talk) 14:30, 16 June 2015 (UTC)

• Rename the test, this is only looking for duplicates
• The user question template should be used
• Are there any other clever things we could do for automation, like comparing headers?
• We'll need a help text for the users to know when the same title is and when it is not appropriate.

Wilco Fiers (talk) 13:57, 7 May 2015 (UTC)

• Use if / else structure
• Add a back to Success Criterion link on the test case page.
• Assumption section is missing.
• The test properties table has a different layout than other test cases.
• The layout of the manual steps are not in line with other test cases. Use Template:UserInput.

Twan van Houtum (talk) 11:39, 7 May 2015 (UTC)

• Step 1 duplicates test SC2-4-2-title. Suggest do ignore pages with zero or more than one title element.
• Step 2: There is no return value, if all titles are different. Suggest to split into two steps. An automatic step "If collection is 5 pages and two page titles are identical, or if collection is more than 5 pages and 20% of page titles are identical." and the manual step, which is already there.

--Annika Nietzio (talk) 14:59, 29 January 2015 (UTC)

http://lists.w3.org/Archives/Public/public-auto-wcag/2014Oct/0008.html by KHS

I am sure that the 50% measure was discussed in the group, trying to address web applications such as email, but still that seems too high in my mind. I also think a manual prompt needs to be included for this test.

My suggestion would be to lower the threshold to 10%, and then add this manual check. My suggestion for an manual check after the first automated one FAILED would be a prompt that says:

"Check the repeated non-unique title instances. Are the repeated page titles are due to a limitation of a web application functional units (i.e., email inbox, email drafts, search results, and <need other examples here>)?"

If yes, do not fail.

If no, fail.

I agree that this is not a perfect suggestion - mbut I am hoping to spark some additional comment on this one.......:-)

• I'm not sure about selecting something across multiple pages. I'd like to have a group discussion on this Wilco Fiers (talk) 08:53, 9 October 2014 (UTC)
• The selector should select the root element, not the head. Wilco Fiers (talk) 08:53, 9 October 2014 (UTC)
• I don't know of any way to automate finding out what a 'collection' of pages would be. Not even sure there is one. We had a discussion on this before and couldn't quite figure it out. Perhaps we could try our idea of looking at all pages linked to from the current page? Wilco Fiers (talk) 08:53, 9 October 2014 (UTC)

• The assumptions say "If there are several title elements for a single webpage, this test will only check the first." This is a different behaviour then in test SC2-4-2-title, which fails if serveral title-elements are present. The assumption of the two tests should be aligned. --Annika Nietzio (talk) 09:43, 9 October 2014 (UTC)
• I like Wilco's suggestion to apply the test on all linked pages. This solves the sampling problem, in case the test is applied to a single page. For benchmarking a whole site, Wilco's suggestion might cause performance problems. But maybe this could be solved with an intelligent caching approach. --Annika Nietzio (talk) 09:43, 9 October 2014 (UTC)
• I agree with Katie that 50% is too high a threshold. --Annika Nietzio (talk) 09:43, 9 October 2014 (UTC)

Ready for review

Closed issues