Meeting minutes
<maryjom> Agenda: Go over Janina's proposed changes
<maryjom> https://
Go over Janina's proposed changes
<maryjom> https://
janina: I did make my updates: how text interfaces are realized; maybe we should talk about utf instead of ascii. Controversial? voice interfaces, chat API
… I rewrote a fair amount, mainly to say these interfaces have seen a resurgence e.g. in programming, revision control
<maryjom> https://
matatk: The update looks fine, Last call we discounted chat-app/ML interfaces, and that's not there as expected
maryjom: Anything you still want to see added or changed?
janina: maybe where ASCII talk about Unicode
jasonjgw: add a parenthetical about UTF8 ?
maryjom: ok let's take some time and comment the PR, okay?
janina: sounds good
matatk: What GitHub calls "reviews" useful in the web, are not accessible in the command line. So instead propose normal comments in the thread
agreed
Go over the spreadsheet
maryjom: Matt and Jason's responses are there. Mary Jo highlighted differences she noticed
jasonjgw: wonders if and where disagreements might be substantive, to focus on those
maryjom: yes that should be the focus, let's go from the top
matatk: made a point of doing it independently of Jason's, then afterwards compared. Apparent differences might come from different ideas of scope.
… Was thinking about terminal emulator, modern ones, new features, might not be accessible
jasonjgw: good to keep modern changes here
maryjom: non text content. Bitmap images or constructed of characters?
jasonjgw: not clear that emulators produce that on their own independent of content they're presenting, could be cases
matatk: I was also on the fence. I can type imgcat and an image appears. Effect on screen reader?
… also strings of garbage text
janina: that matters
… change no to yes?
jasonjgw: unclear, if operating system handles it, depends where it falls. Maybe on terminal emulators, yes on content
maryjom: changed answers to 'maybe' in the spreadsheet
jasonjgw: imageMagic can run, separate window?
… don't do audio, but use system audio
janina: let's not assume single audio device
mitch11: Two questions: 1. Comment about garbage text - referring to ASCII art?
matatk: yes ASCII art; also proprietary escape codes to represent the image
jasonjgw: screen readers get characters after escape codes have been processed, but don't know what that looks like for graphical image
mitch11: 2. About the spreadsheet: what did we mean by "problematic"? Difficult to do, or can't be applied?
jasonjgw: If I can't think of obvious case where terminal based program or terminal emulator could fail a criterion, then not problematic
… if it could fail, then marking yes problematic, and not too far-fetched; person evaluating can make a determination
… Next we go through all criteria, or just the significant differences?
maryjom: AAA is less pressing, so not immediately
… can go into concepts: graphics; programmatic information
maryjom: how about we continue in top down order: skipping the ones with already agreement between jason and matthew
matatk: meaningful sequence is a good start, maybe we shouldn't expect an accessibility tree
… proposal has been heard we don't need inserting semantics, but fact is related escape codes exist
… with attention on what a person needs to do their job
… have seen printing output in columns, if wide enough display, question of recognizing tabular structures, probably not that common, and need to draw a line
jasonjgw: traditional way is application has control in it where it simplifies the terminal presentation, sometimes so a program can process output,
… sometimes an accessibility consideration, like true cursor location
… so wondering to what extent problems are configurable
… diff between character cell presentation, versus what's presented in the terminal
matatk: was interpreting it as could these criteria be failed; but could have asked whether the criteria relevant; and not relevant got a "no" response
jasonjgw: same
janina: when got Unix versus DOS. Filename on the right had my ire
… problems having labels above a field, not performative or useful
… there's a standard extension to flag the fact that hey there's a text based interface, but many are problematic
… point is consider not making those accessible, but if they're just alternatives to graphical environment where objects have properties
… focus on there is an option that is accessible
mitch11: I like the point that GUI could be an accessible alternative for a TUI. That seems to be a decision for an evaluator, but not a reason for us to change how we evaluate a given text interface, if we evaluate it.
… To the point that text interfaces have got worse for accessibility: the old ways of doing things (e.g. line-wise, as Janina mentioned) are still there. We should keep our high expectations of accessibility, as nothing has been depracated.
jasonjgw: can pipe it somewhere, often have options, even space separated fields could lead to programmatically determined structure; might have to do it per tool but then it's reliable
janina: yes with learning curve
jasonjgw: fixed width coluimns
… emacs can identify headings, terminal escape sequences, man page
… more can be programmatically detrermined than might at first seem
mitch11: A nit about a text file being programatically determined - maybe we should say "programatically, or in text, determined". We could point out that the "or in text" is a substitution.
jasonjgw: Not sure we can successfully look at text applications with a WCAG lens, we can often make things fit, but it can take effort.
janina: and database back ends are structured
mitch11: Think we should make the distinction that examining the text is one way to subsitute in the info that we might otherwse gain programatically.
maryjo: concerned if we allow "or in text" then gives free rein to go backwards, where programmatic is also available, it would need to be scoped narrowly if we said it
janina: agree
jasonjgw: worried that these should be limited to edge cases
maryjom: like closed functionality, could be something similarly narrow to command lines
mitch11: Agree; we should use language in this specific section that is appropriate, but doesn't redefine WCAG terms.
maryjom: task for the week: comments on these, email conversations
matatk: should we have fixed set of responses in the spreadsheet? e.g. yes it's a concern that can happen, no it can't, or not relevant
maryjom: good, jot down such thoughts in comments
… when meet again?
… focus on the criteria down to 2.1.1 Keyboard, meet in a week
agreed