Silver Task Force & Community Group

28 October 2022


janina, jeanne, Lauriat, Makoto, maryjom, Rachael, sarahhorton, SuzanneTaylor
jeanne, Shawn

Meeting minutes

<jeanne> queue:

<jeanne> Scribe List

<Lauriat> https://www.w3.org/WAI/GL/task-forces/silver/wiki/Scribe_List

Preview of Tuesday AG WG call

<jeanne> survey

Rachael: Plan to spend an hour on the Conformance Models survey

<Rachael> The survey takes some time but is really important because it will help us set the direction we move forward in

Review asynchronously drafted user needs

jeanne: Recap - we are working on a project from TPAC to explore whether outcomes can be written as user needs

jeanne: We have working on alternative text. Made a list of flows with examples where images need alt text.

jeanne: There are volunteers who are doing a more in-depth analysis of the flows

<jeanne> Alt Text Writing OUtcomes as User Needs

SuzanneTaylor: Took decorative images to start with. There will be more things for "find images for children" and decorative images.

SuzanneTaylor: Teens were happy with very brief alt text, even though they were decorative images. While other users might not want to have decorative images identified.

SuzanneTaylor: If there are images that are in a textbook, it was hard to tell if a decorative image or an image without alt text was important to learn about.

SuzanneTaylor: Users with anxiety may also be affected as it isn't obvious whether the image is truly decorative or is something that they really need to know about.

SuzanneTaylor: Need to go through the list of functional needs. Will go through some of the cognitive user needs so that decorative images don't cause distraction, fear, or anxiety.

sarahhorton: For images, on links and buttons users need to distinguish them from other images. Need to understand the function of those images.

sarahhorton: Programmatic labels are needed for functional images (through alt attributes) and conveys the same information as the functional use of the element.

janina: We may find that some applications will be important for users that aren't using AT. WAI Adapt can help us explore implications of that.

jeanne: Under programmatic labels we can make a note that not all technologies have access to the accessibility tree.

janina: It might be cool if I could hide the functional ones while I tried to understand the content - and can be assured they won't accidentally trigger something.

jeanne: I added some resources in looking at this and added to the document for further research. Examples of flows and images section.

jeanne: There are common and unique needs, made a list of needs for graphic novels.

jeanne: Read through the bullets regarding Graphic Novels under common and unique needs.

jeanne: There is a comic book markup language, written in 2014

jeanne: There's more to graphic novels, but this will help start talking about granularity - the hard problem to solve.

jeanne: What's the granularity of information that should be provided, based on the type of content. What guidance should we give?

discussion of granularity depth for user needs based on the examples

sarahhorton: There is a tension between what people need and how to get that to them in a standardized context.

sarahhorton: What's needed and how you do it is something that may need to adapt based on functional needs.

<SuzanneTaylor> +1 on left shift to guidance on how to design

sarahhorton: Need to approach from a design perspective and give guidance to designers so they know how to do that well.

<janina> We need to know that in code; so that software can respond based on a user's configured settings

sarahhorton: Identify functional from non-functional images, and ensure it is recognizable and distinguishable from other things.

sarahhorton: The next question is how to make that testable.

jeanne: I added to the Notes on the Process section the three areas of tension.

sarahhorton: May want to add to those notes about design decisions (did some editing of that in the document).

janina: Maybe clip from the transcript what was captured, as it was said well. We may need additional markup to define this and identify it semantically.

janina: Looks like we need enhancements to existing markup to support more programmatic functionality

jeanne: Assuming the problem we're defining is what is the level of granularity that we should be talking about and are multiple levels needed?

jeanne: Is there a user need in the common list we can apply to all flows and would we want user needs for the unique needs at a different level?

jeanne: ... and putting more unique needs as the more advanced things we ask people to do?

Lauriat: Rather than going into details of animated images, image buttons, etc. and instead express the user need for all of these things and how to approach it.

Lauriat: Have some examples that would be lower granularity - showing the range of user needs. e.g. An image complementing an article.

Lauriat: Another is an image button as part of the toolbar, and helping the user understand that there is an interaction that is needed her. There may be a user need for more verbosity.

Lauriat: And help people understand the user needs in the context of what is their goal/flow of what they are trying to do.

<Zakim> Lauriat, you wanted to suggest a high-ish level, with lower-level illustrative examples

<Makoto> https://docs.google.com/document/d/1ExW-rY7K00un18z3Z243LvmznQ2YRCHNKAO6TDHKIiM/edit?usp=sharing

<jeanne2> Makoto: I did an analysis of informative images

<jeanne2> ... I did it in a separate document

<jeanne2> ... I was part of the original group that worked on alt-text

Makoto: We can use several of these images with graphic examples.

jeanne: there's all the different types of images in this analysis.

jeanne: Could we be looking at the user needs at the top level of sorting Makoto did and have additional guidance at the lower levels.

Lauriat: We could have a fairly generic statement for the top, top level (one level) and then use the more detailed user needs to show the more unique needs.

SuzanneTaylor: Like the detailed guidance to be examples. If we make it prescriptive, it may not work for all cases.

<janina> +1 to Suzanne

SuzanneTaylor: We don't want to have things too prescriptive, as different users may require different levels of description

janina: We have to be careful that if there needs to be updates to something like the HTML spec that we capture all proposed changes in one pass.

jeanne2: Going back to the tensions Sara was talking about, what would be examples of the different tensions and what we're discussions and what the granularity should be.

janina: This is something we may consider putting out in a draft and asking for public input, so we're not a small group grappling with these things.

janina: Then we can get that input from real users.

Rachael: The next draft is planned to be in January. The next draft will have a couple of conformance options.

jeanne2: We want some examples to go to AG WG in a week or two of what we think user needs could be in the current WCAG 3.0 structure.

jeanne2: For images, here are some examples of how you would write a requirement so AGWG can discuss whether or not they want to do it that way.

sarahhorton: Is this instead of outcomes?

<Lauriat> https://docs.google.com/spreadsheets/d/1tXIaJxQCyuOTkvLcJR5QgTN13h6dIgtUlE7H9CUCvew/edit#gid=0

jeanne2: Yes, this was discussed at TPAC and they wanted a sub-group to come back with examples. Instead of a sub-group, we are working on that in these Friday meetings.

Lauriat: My thinking for next steps and for each functional need noted, write an outcome for each. Have a couple of lower-level illustrate examples for how to test it.

jeanne2: I think Makoto has outcomes.

Makoto: I had list items of examples of 7 types of images. Not necessarily a fully comprehensive list.

jeanne2: The original text alternatives only had one outcome. We need outcomes for each one and user needs for each one. So we can give an example to AG WG of different user needs and accompanying outcomes with informative examples.

jeanne2: We can do this for each type of image. Is this a good model?

Lauriat: Makes sense.

jeanne2: Homework: Take a look at Makoto's list and take a pass at writing an outcome and accompanying user need.

Minutes manually created (not a transcript), formatted by scribe.perl version 192 (Tue Jun 28 16:55:30 2022 UTC).


Succeeded: s/Chuck, you wanted to say this is our first observation :-)//

Succeeded 4 times: s/sarahhorton:/SuzanneTaylor:/g

Succeeded: s/sarahhorton:/SuzanneTaylor:/

Succeeded: s/prescriptive, but/prescriptive, and/

Succeeded: s/prescriptive, and/prescriptive, it/

Maybe present: jeanne2