Silver Reliability

05 Apr 2022




<scribe> scribe: dmontalvo


JS: Probably good to talk about the exercise last week in AGWG to see how it will affect our work

FS: Not sure about what they want to do with the results

JS: They plan to do this for each SC. I want to make sure we agree on this, especially where test outcomes should go
... The exercise was about categorizing SCs to move them to WCAG3

<Francis_Storr> Protocols and Scoping Proposed Next Steps Google presentation: https://docs.google.com/presentation/d/1b5xHQWBzoYdKp7BfPgIUBCpz-yaDOx_kSq_HlQxcFh0/edit#slide=id.p

Wilco: We ended in a similar direction, with us focusing only on text to speech for now
... I think the exercise was fine

Pronunciation of Text


<Work on bullet point 11 in the last page>

Wilco: If we are talking blocks, tihs won't be an exception. But this example is somewhat intentional, maybe we should consider these things

JS: Somebody can copy a block of text from a German website and put it in a Spanish website without changing the lang value

FS: If we change the example to a block of text that would be a valid thing to do

JS: Shall we do this for when declaring the language is not necessary?

FS: I think we should
... Unless the word has become part of the natural language, like in WCAG2

JS: I still think even if it is part of the natural language and it is wrong, we should address that

<jeanne> JS: An incorrect language code should always be wrong, even if they don't have to declare the code

Daniel: This seems like a wrong use of a lang value, probably we should pint that out

FS: We don't have failures (critical errors in here)

JS: Probably that would be too much

WF: I would say this is wrong, but if so that would add an extra level of complexity

JS: You would need to go through the HTML and see if the value is correct

WF: What if you need to test in a programming language where you don't have access to the underlying code?

JS: That would require a screen reader text, would not know if this is a site bug or a screen reader bug
... This seems like this is a cumulative problem. If there is one, not a big issue, if there are more, then it is a significant issue

WF: Maybe we should be oK with that happening incidentally
... If we are looking at the outcome, it should not matter who or what is creating the problem
... This case gets into intention. In a legal context intention matters.
... If someone is injured because of an accident, that is different from that happening because of a specific intention

JS: Interesting idea, how some of the use cases would look like?

WF: It is different if human decisions are making the outcome wrong than if machine decisions are making it wrong
... Automating this process will throw some errors. IF these errors are created by humans, that is a different thing
... Does that matter?

JS: This encourage innovation for language detection and it discourage human tweaking of lang values that may reveal lack of competency

WF: If you do it automatically we have a greater tolerance for errors

Daniel: How would a tester know if a passage has been produced automatically or by a human?

FS: That would create a huge burden on the tester

WF: I can understand why we want to be stricter in that case, but it is going to be much harder

FS: This would imply having a look at particular CMS and plug-in versions

WF: The conclusion is we are not failing individual words, even if they have a lang value that is wrong, because that would take a significant amount of work

JS: I would reconsider my position here, as this would preclude automation

<updates to the document based on previous discussion>

Bullet number 12

Daniel: This one is unclear for me
... The pronunciation rules can vary depending on the speech synthesizer
... Pronunciation rules can be defined at multiple levels: speech synthesizer, screen reader, user ...

<changes to document based on discussions>

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes manually created (not a transcript), formatted by David Booth's scribe.perl version 1.200 (CVS log)
$Date: 2022/04/05 14:00:26 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision VERSION of 2020-12-31
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

WARNING: No "Present: ... " found!
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy
        <amy> Present+

Found Scribe: dmontalvo
Inferring ScribeNick: dmontalvo

WARNING: No date found!  Assuming today.  (Hint: Specify
the W3C IRC log URL, and the date will be determined from that.)
Or specify the date like this:
<dbooth> Date: 12 Sep 2002

People with action items: 

WARNING: IRC log location not specified!  (You can ignore this 
warning if you do not want the generated minutes to contain 
a link to the original IRC log.)

[End of scribe.perl diagnostic output]