<trackbot> Date: 22 May 2015

<richardschwerdtfeger> meeting: SVG A11Y Task Force

<richardschwerdtfeger> scribe: Rich

Timelines and Prioritization

<richardschwerdtfeger> Rich: It sounded like SVG2 to go to CR at the end of they year to first quarter

<richardschwerdtfeger> Amelia: yes, after TPAC

<richardschwerdtfeger> scribe: Jason

<jasonjgw> Rich suggests we develop a basic graphics model as a high priority. We need to continue to work on navigation, noting that tabindex is a good start.

<jasonjgw> Amelia: Ideally, I would like to see more detailed navigation work in the core spec.

<jasonjgw> Fred asks about the two-implementation requirement; Rich confirms this and suggests that anything not mapped to role=none should be focusable.

<jasonjgw> Rich: this way we could focus almost anything that might be wanted.

<fesch> +1 to foucusable for anything but none

<jasonjgw> Fred agrees. Amelia suggests distinguishing focusable from tab sequence - not everything should be in the tab navigation order.

<jasonjgw> Rich: these elements would be effectively tabindex=-1.

<jasonjgw> Amelia agrees to take these issues to the SVG working group over the next few weeks.

<jasonjgw> Rich notes that under this proposal we wouldn't have to add tabindex=-1 to all of these elements, and it opens the way to further work on navigation.

<richardschwerdtfeger> jason: I think that is a good idea.

<richardschwerdtfeger> jason: tabindex tends to go with interactive elements.

<richardschwerdtfeger> jason: What is focusable should be a distinct manner

<richardschwerdtfeger> jason: I am comfortable with the proposal regarding making these non “none” elements focusable

<richardschwerdtfeger> jason: Navigation similar to what Rich proposed moves us in the right direction.

<jasonjgw> Rich notes that we need editors to work on the separate document on semantics. This would be an ARIA graphics module. Rich notes that we need basic primitives - for shapes, allowing a fall-back that makes sense.

<jasonjgw> We would then add further roles for charts. Rich suggests we need to add some of these to SVG 2.

<richardschwerdtfeger> http://rawgit.com/w3c/aria/master/aria/dpub.html

<jasonjgw> Rich notes the digital publishing ARIA extension taxonomy, referring to it as a model for what needs to be created in SVG.

<jasonjgw> Amelia regards this as achievable.

<jasonjgw> Fred expresses willingness to edit the document.

<jasonjgw> Fred notes that Charles may be interested in serving as an editor and suggests deciding this over the next few weeks.

<jasonjgw> Note that Amelia has agreed to serve as an editor.

<jasonjgw> Fred wanted feedback on the CSSUI work and comments on the use of role=text and the application of ARIA for purposes other than assistive technologies.

Use of role=

<jasonjgw> Amelia summarizes discussion in the SVG working group. There was more support for the idea of having a means of recognizing images of text as text.

Role=“text” for other mainstream features copy/paste, search: etc.

<jasonjgw> There was no agreement as to whether this should be done by way of new SVG elements or ARIA roles.

<jasonjgw> It was resolved to develop a clearer set of proposals/options.

<jasonjgw> Rich clarifies that the use of roles for other purposes wasn't raised as an issue in the SVG working group discussion.

<jasonjgw> Responding to Fred, Rich notes the issue raised on the list of whether SWITCH could be applied to TITLE and DESC.

<jasonjgw> Amelia clarifies that it apparently can be used but not in the ways one might anticipate. There's an issue to be addressed regarding TITLE as a child of SWITCH.

<jasonjgw> Fred clarifies.

<jasonjgw> Child as a title of SWITCH never appears as a tooltip.

<jasonjgw> Rich notes we need an additional step in the name/description computation. This work is required in the SVG spec.

<jasonjgw> Amelia is planning to prepare edits for Rich to review upon his return.

<jasonjgw> Amelia notes the issues raised on the mailing list regarding VIEW.

<AmeliaBR> https://lists.w3.org/Archives/Public/public-svg-a11y/2015May/0044.html

<jasonjgw> Amelia suggests behavior that would treat a link to the VIEW element as though it were a change on the SVG element.

<jasonjgw> Amelia offers to create examples of views and to encourage experimentation by participants on this task force.

<jasonjgw> Amelia notes two major use cases: zooming into parts of a large graphic, and identifying the correct icon (e.g., for screen readers) when a number of icons occur in an embedded SVG document.

<jasonjgw> Amelia notes that VIEW is very different from anything in HTML.

<jasonjgw> Fred suggests it's somewhat similar to popup dialogs.

<jasonjgw> Fred also notes that point of regard is problematic here; Amelia agrees, noting that the VIEW itself isn't the correct destination for point of regard.

<jasonjgw> Amelia suggests the target attribute should be an implicit aria-flowto.

<jasonjgw> Rich notes that aria-flowto is well supported under MS-Windows but he needs to confirm support under VoiceOver.

<jasonjgw> Fred notes potential applications of view in technical drawings where there is a plan view distinct from a side view.

<jasonjgw> Amelia notes that there are numerous use cases.

<jasonjgw> Fred notes further use cases in data visualization applications.

<jasonjgw> Rich notes that aria-flowto can take multiple idrefs, enabling the user to choose the destination.

<jasonjgw> Rich notes that the reverse relationship (from the flowto destination back to the node with the flowto) is implicitly handled by the user agent and not by the author in order to avoid authroing bugs.

<jasonjgw> Amelia notes that not everything which has an ID will have an accessible name and hence name calculation can be problematic in SVG.

<jasonjgw> aFred notes a limitation of aria-flowto: it is directional and cannot represent undirected relations.

<jasonjgw> Amelia notes that these can be represented as a pair of flowto relations, one in each direction.

<jasonjgw> Fred notes a further practical issue with screen reader implementations of flowto - that the set of usable/nmemonic key bindgs is quickly exhausted as the number of flowto relations increases.

<richardschwerdtfeger> http://www.w3.org/TR/core-aam-1.1/#mapping_additional_relations_reverse_relations

<AmeliaBR> scribe: Amelia

<AmeliaBR> Jason emphasizes the importance of having a way to designate undirected connector relationships, but suggests that it can be simplified because most graphs are either directed or undirected

<AmeliaBR> Fred suggests this might not be true. E.g., in an org chart you could have sideways connectors between equal employees as well as directed connectors from supervisors

<AmeliaBR> Jason suggests that navigation could still show all the flow options somehow

<AmeliaBR> Fred points out that this isn't how it currently works: e.g., JAWS uses different key bindings to reveal flowto versus flow-from relationships

<AmeliaBR> Jason gives examples of where this is a limitation: wanting to find how many edges are connected to a vertex within a network graph

<AmeliaBR> Fred notes that, in a model he created for sighted users, he presented both types of relationships at the same time, but grouped into flowto and then flow-from

<AmeliaBR> Jason notes another issue, for navigation models, is what to do with unconnected graphs, where there are separate groups with no connections between them; you need a separate navigation method

<jasonjgw> Amelia suggests we need to decide upon priorities for the SVG 2 core spec. It appears from what has been said that establishing a basic graphics taxonomy is a priority, as well as updating the accessibility API mappings.

<jasonjgw> Fred notes this as a priority - establishing a basic taxonomy.

<jasonjgw> Fred proposes that we start working on it next week, raising the question of whether we start from where we left off in earlier discussion.

<jasonjgw> Amelia suggests starting from existing work, offering to move it into the wiki.

<jasonjgw> Amelia agrees to move her work on this to the wiki.

<jasonjgw> Fred notes that this will then be the basis for discussion.

<jasonjgw> Amelia proposes to continue work on the accessibility API mapping issues.

<jasonjgw> It is agreed that animation raises substantive issues.

<jasonjgw> Amelia notes that animation can be used to hide/reveal graphical components of the SVG document.

<jasonjgw> Amelia suggests treating animation similarly to changing the CSS DISPLAY property on an element, and will come up with options for addressing the issue.

<richardschwerdtfeger> ACTION: Ameila meet with SVG working group regarding focusable elements vs. tabindex [recorded in http://www.w3.org/2015/05/22-svg-a11y-minutes.html#action01]

<trackbot> Error finding 'Ameila'. You can review and register nicknames at <https://www.w3.org/WAI/PF/Group/track/users>.

<jasonjgw> Fred notes his discussion comments on Amelia's work; Amelia offers to move them into the wiki.

<jasonjgw> Amelia plans to read Fred's comments.

This is scribe.perl Revision: 1.140  of Date: 2014-11-06 18:16:30  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/Amila/Amelia: Ideally, I /
Succeeded: s/model her/model he/
Found Scribe: Rich
Found Scribe: Jason
Found Scribe: Amelia

WARNING: 0 scribe lines found (out of 102 total lines.)
Are you sure you specified a correct ScribeNick?

Scribes: Rich, Jason, Amelia
Present: Fred_Esch AmeliaBR Rich Jason_White
Found Date: 22 May 2015
Guessing minutes URL: http://www.w3.org/2015/05/22-svg-a11y-minutes.html
People with action items: ameila