W3C

– DRAFT –
WCAG2ICT terminal applications

21 April 2023

Attendees

Present
dmontalvo, janina, jasonjgw, maryjom, matatk, Mitch
Regrets
-
Chair
Mary Jo
Scribe
dmontalvo

Meeting minutes

First thoughts on terminal applications

Janina: JAson was an editor in the previous document
… The two sections that are there are the result of that work
… I think we want to talk about how much of that needs to be updated
… I think there is also the topic of whether they are additional SC that pertain to this

MJ: A third one is whether there are problems or things that might not make sense to apply in this context given new SC 2.1 and 2.2

Janina: That's on my mind too. Haven't done that work yet

Jason: I think that is an importat step to try
… Ther's nothing said about the innterpretation of the existing SCs in that context, it's just about the broad principles
… If there is anything that should be said that is more specific, we should do it
… We should take issues that are not related to WCAG elsewhere, not clear where though

Janina: I'd say that's APA's responsibility
… If it's not WCAG, it should be APA

Matthew: Sounds like it's a good fit

Jason: IF there are web based terminals now that's another thing
… The appendix says something about the accessibility of the application that is different from the accessibility of the terminal itself

MJ: If we find gaps that need to be resolve, we are not the group to act on them at this point. We are now focusing on how WCAG apply to this context

Janina: I agree, thank you for saying it
… I do think we need to look at the whole range to decide what needs and does not need to be coveredwhawt

What needs updating?

Janina: Command line users were treated as fossil. Glad that this was taken up at the beginning
… These are no longer historical, they are now under active development, there are many "new terminal" concepts these days
… There ar etechincal things, like references to ASCII
… Nowadays you can do much broader of a charset
… There are Coga implication
… There is now new concepts that have reinvented the terminal
… Latency matters, but I am not sure if that's on WCAG
… By databasing the interface they made it possible to build a terminal interface that is context-sensitive
… Massively improve the user experience, for example checking a man page
… I talked to Lissa about it
… They may benefit from use of gh-cli
… The reason why terminal is revived is because it is know much more context-sensitive

MJ: What's the WCAG context for this?

Janina: Maybe not calling it historical anymore. That framing is not accurate

Mit: I love the background. What's the scope? Things that run in terminal applications is clear to me

Either command line or full screen text-based
… When we say just text-based I am not so clear
… I thikn that is a rather different thing

MJ: I was thinking of terminal applications

Janina: All we had at the time was the search command
… I was putting it out as a question

Jason: That is an interesting open question, there are tools that let you run these things in a termianal
… I know the existing text tries to distinguish the accessibility of the programs and that of what is running those programs
… I think it is definitely true that automatic shell completion is now much more significant these days
… Not sure how much these fit into WCAG though

Matthew: I think it'd be a good idea for us to define this distinction in the document
… First attempt was Text/command line
… These were basically things that run in a terminal
… IF we are thinking of applications where text is he user interface we should make it clear too
… That may be part of the revision
… I know you can't directly apply 1.4.10 to terminal, but there is an aknowledgment that people may have many terminal windows opened
… There should be the ability for these to adapt to different viewports

Jason: The scope of this work is for non-web ICT,
… What if there is a website tool that is runinng a terminal that is connected to a secure shell that's also connected to another termianal
… Ultimately the web should be accessible but there are implication in the other parts as well
… so we should be looking at the accessibility of those as well
… Certain commands have built-in options to pipe the output in a way that it matches the given window size

Mitch: I'd choose that the scope be things that run in a terminal
… We are talking about content and applications, so it make sense to talk about them as if they are just ext, no controls or other ways of interaction
… We may need to also qualify the other ways of interaction to be clear about scope
… I don't know to what extent we should address web-based terminals
… These are webs that would fall under WCAG, but they are also runnig "just text" so it is difficult to make that distinction
… There are differences in screen reader behavior between native web and native terminal, how much of these should be covered?

Janina: That is a good principle, and a good waay to distinguish things. Every terminal app does not have controls you can click/press to activate
… For all the other switchers it all depends on your environment
… I am going to look if we have something on WCAG that kills the way some manufacturers do terminals - redrawing the whole screen on every keypress
… Probably we don't have it covered
… Some WCAG application is the ability to select the content that is being displayed
… How do you do your system logs as an engineer? I have only done it through the command line, I know that there are graph tools etc

<Zakim> matatk, you wanted to discuss 1.1.1, 1.4.1, 2.3.1

Janina: This exist, so I want us to make sure we think of those

Matt: 1.1.1 and 1.4.1 I mentioned those in the notes I sent earlier
… Flashing may be an issue. If you have a huge file even if it is just ext, the screen would flash. Not sure to which degree that is dangerous, but we need to look at it to see if we can include wording for the terminals to have the ability to disable this behavior

Jason: This is a good example of it.
… Basic accessibility requirement is that everything be operated via keyboard. Some of thesegraphical terminals do not meet this
… If they do, not sure how accessible these will be for a screen reader user
… Selecting, copying, and pasting text is also an issue

MJ: We will have to lookat terminal applications to check which of these SC are problematic

Jason: Some may be already covered by existing SC

MJ: I am talking about the non-web terminal applications. Web-based should meet WCAG

Jason: They would have to satisfy if the web needs to comply to policies and such
… That does not mean there's no gaps in these types of applications

<matatk> ack ;-)

Jason: Those gaps are probably out of scope for this discussion though.

MJ: Not saying this is not an issue, just saying that there's WCAG for that
… IF there is a missing criteria that should be outside

Janina: We mean WCAG3, then
… Which will not happen in a few years

Matt: I've seen a lot of developments that could be fantastic if they were exposed in an accessible way
… I think we would still count a pseudo control like a text box maid of icon symbols as textarea
… Sometimes we exempt a platform because there is no accessibility support
… In mobile apps you could not do headings, now you can
… IF you have a web-based terminal and you run an unmodified command terminal it is not going to take advantage of the platform, that should be noted as a caveat

Mitch: Back to the control question, I corrected scribe as I think it was not reflecting what Janina said

Janina: The controls may not be constructed into widgets

Mitch: This reminds me of 1.3.1 language but 4.1.2 does not have that language

Next Steps

MJ: How do we want to work? What are next steps?
… It sounds like there are some smaller content changes that need to happen, we should be clearer about the scope
… We may need to make some technical changes as well
… The appendix goes into further detail but it does not get into the analysis of each SC

Matthew: What we have discussing could be considered a patch for terminal applications, sounds like reasonable next steps
… Happy to help

JAnina: I would volunteer to take a pass at the summary section
… On the appendix I think we should be able to say Yes, No, Maybe, for each of them

Jason: I'd volunteer to participate

Mitch: Me too

Janina: We cannot change anything it says. These changes need to be WCAG3

MJ: How should we work on tihs?

Janina: Happy to create an ASCII spreadsheet

MJ: I can then turn this into a spreadsheet

Matt: I loved Mitchell's calling terminal applications "user agents"
… I think we should include this somewhere

MJ: When should we get back together?

Janina: Two weeks from today would be 5 May

Janina: It works for me

Mitch: Sounds good to me

MJ: We have tasks to do, we can email between us, Janina, when you have the spreadsheet please send it my way, let's try to have something together a couple of days before the meeting.
… Everyone please by 3 May get things to me

Minutes manually created (not a transcript), formatted by scribe.perl version 210 (Wed Jan 11 19:21:32 2023 UTC).

Diagnostics

Succeeded: s/manufacturers do terminals/manufacturers do terminals - redrawing the whole screen on every keypress/

Succeeded: s/Not every terminal app has controls/Every terminal app does not have controls/

Succeeded: s/Chat GPT is an example of a "new terminal" concept, and there are many others/there are many "new terminal" concepts these days/

Succeeded: s/thes /these/

Succeeded: s/create and/create an/

Succeeded: s/email ourselves/email between us/

Maybe present: Jason, Matt, Matthew, Mit, MJ

All speakers: Janina, Jason, Matt, Matthew, Mit, Mitch, MJ

Active on IRC: dmontalvo, janina, jasonjgw, maryjom, matatk, mitch11