Accessibility Guidelines Working Group Teleconference

20 Apr 2017

See also: IRC log


allanj, Laura, AWK, Glenda, Greg_Lowney, ScottM, JF, Katie_Haritos-Shea, MichaelC, KimD, jasonjgw, dboudreau, Lauriat, shwetank, Makoto, Melanie_Philipp, Detlev, Rachael, Wayne, MikeGower, marcjohlic, Bruce_Bailey, Jake, steverep, jon_avila, adam_solomon, Elledge, Mike_Pluke, kirkwood, Kathy, LisaSeeman, Michael, seeman, DAVID, Shawn, Joshue108, JohnRochford



I can scribe, let me go get my scribe cheat notes

<scribe> SCRIBE: Glenda

<JF> scribe: glenda

OPEN item 1

Joshue: struggling with all the different wordings. Main stumbiling block seems to be around font.

<david-macdonald> Is wayne on call

<Joshue108> https://www.w3.org/2002/09/wbs/35422/AdaptingTextSurvey/results

Alastair: discussioin with Patrick was around font aspect. Patrick objects to using font in this SC.

<JF> +1 to Josh

Joshue: is including font muddying the waters?

Alastrair: we are trying to achieve overriding text styles. Most things are really spacing related. We often achieve this by changing font family.

Alastair: If we can specify a way to ignore font icons, the problem goes away.

<Joshue108> +1 to JF

JF: if user changes all fonts on the page to one of the user’s choses, why is this a problem?

<allanj> +1

Wayne: 2.0 did not pay a lot of attention to semantics that happen inline. When someone puts an icon font, or a may symbol, the actual way that thing looks, has meaning.
... We need to add some exceptions for font icons that provide specific meaning that are not availalbe in another font, we have to preserver that meaning.

<laura> The principle is that if you override the font, the issues appear. E.g. font-icons disappear. Same for colors, the act of changing them at all will highlight most or all issues people would get from changing them to their preferred colors.

<david-macdonald> for those who are not talking please mute... we have crackling on line

Joshue: looking at this SC objectively, there is very little that points to semantic nature of some font usage. So, that to me suggest that there should be an addition to this SC about semantic aspect to iconography or for example a mathematical operator of formula will need to be preserved.

<Zakim> Greg, you wanted to say I'm somewhat concerned that the choice of two bullets can lead to sites requiring people having to go through a password reset procedure every time they

<Greg> 1. Major browsers let the user override author-specified fonts and/or colors, an accessibility feature that is easy to use and well documented. When an author makes their page intentionally incompatible with that feature, I don’t think we should allow them to claim to be sufficiently accessible.

<Greg> Note that marking up use of icon fonts would not fix this problem.

<Greg> 2. Font is a major stumbling block, but another issue is whether we require compatibility with a wide range of appearance schemes, or just two (one default and one alternate).

<Zakim> AWK, you wanted to say that my takeaway is that icon fonts can be used, but need to be used properly. The outstanding issue is that we need a metric to make sure people don't need

<LisaSeeman> but it meets the humen testable bar

<AWK> https://rawgit.com/w3c/wcag21/AWK_adapting-text/guidelines/#adapting-text

AWK: having trouble with the testability of the font bullet point. There are problems that need to be addressed, but testability is problematic for font. I think you could always find at least one font that would pass, so I don’t think this has a testable solution yet. Can we take fonts out of this right now (with an editor’s note that this is something we will research and add in at a later version).

<david-macdonald> +1

<alastairc> +1 to Andrew's proposal (note it is missing a close bracket by captions).

<Zakim> Joshue, you wanted to say then there is a semantic component to the use of fonts that in some cases needs to be preserved.

<Joshue108> JOC: Some sample text to cover semantics - Each of the following text styles of the page can be overridden with no loss of essential content, functionality or meaning.

<david-macdonald> @ awk close the bracket (with the exception of images of text and captions,

David: AWK in the short term solved it for me. Looking at Icon fonts the same way we looked at table layout. We don’t recommend you do it, but if you do it, do it in an accessible way.

Jason: could we fix the fallback algorithm so if a character is not found in the replacing font…that the fallback would be to the original author supplied font.

<AWK> We might also need to clarify that the replacement font needs to support the same language. If I replace arial with a korean font I can't expect a good result

<JF> +1

Katie: I like AWK’s idea, of going forward without the font at this time, and doing furuther research on font. Glad to see the energy that is being put into the font solution.

Wayne: In line characher sequences that are part of the natural language like math or physics or icon fonts (like emojis). We cannot write a computer program that can parse out the meaning of an author. Authors will need to add role or identifiication when the font has specific semantic meaning and cannot be replaced with just any font family.

Jason: I still think there is a possibility of falling back to the author supplied font instead of just displaying squares. I do understand that there may be use of font family that relays semantic meeting and would need author supplied semantic markup.

<Zakim> AWK, you wanted to ask about failures related to the color changes

<laura> color issues: graphical backgrounds (single colour or gradients rather than pictures) that make text unreadable if you reverse the colour scheme.

AWK: we still have challenging issues related to font. I would like to address these font issues, but also make progress in this SC in 2.1 as well. What about the color issue? If the end user can change the text color and background color to a single different text color and a single different background color, is that enough.

Wayne: I have been classifying failures. Big failure is on high contrast and background images. When author assumes that various icons are sitting on top of a specific background image color.

<AWK> But the icon isn't text? or is it?

<Zakim> Greg, you wanted to discuss color use cases

<Greg> Use case: user changes to light on dark, but icons are hard-coded dark to contrast with the default background color, and so becomes invisible.

<Greg> Use case: information is conveyed by text color differences (e.g. new items in red, or alternating colors to achieve animation), which disappear when the user overrides author colors.

<Greg> Use case: user want to avoid text colors that conflict with their synaesthesia.

<LisaSeeman> any chance we will get to authetification today?

<alastairc> The plan is to create a testing speadsheet and go through many examples....

<LisaSeeman> it might be better to finish this one as we will not have enough time to finish authification anyway

Josh: if we can step back, can we remove font family at this time. How can we progress this now?

<LisaSeeman> (what a shame we missed that John....)

JF: icon fonts already need an aria-label for meaning for screen reader users, but for sighted users, they don’t have access to this.

<alastairc> so where do we put the semantics for this? ARIA? HTML?

<alastairc> Patrick is objecting to a non-normative technique

JOC: we could have a technique for aria to show semantic of the icon font

David: Paul Adam wrote a plug in to reveal aria labels and make the visible. Wouldn’t a technique be to make aria semantic info be visible for font icons.

<allanj> this could be done with CSS also

<Greg> John, note that any add-in that makes hidden text visible will change line widths; if the page can handle that without breaking, they can also handle changes to font widths.

Lisa: The last thing you want to add is more info on the screen for the low vision user (if it is not needed). I’d like to see personalization solutions. Not good to simply show all the aria labels.
... can we have straw poll and see if we can reach consensus

<Wayne> The ratio of parameters of a box with a specific font like ARIAL, to an arbitrary font X for all boxes with the same permiter.

JOC: can someone explain differences between H & I versus J & K versus L & M https://www.w3.org/WAI/GL/wiki/Issue_78_Options#Proposal_L_text_reads:

<Greg> Big difference: whether the Level A requires testing with a wide range of appearance choices, vs. only two (one default plus one alternate).

JOC and AWK: We will leave this open for now.

RSSAgent, make logs public

trackbot, end meeting

Summary of Action Items

Summary of Resolutions

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.152 (CVS log)
$Date: 2017/04/20 16:29:19 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.152  of Date: 2017/02/06 11:04:15  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00)

Succeeded: s/or a mathematical operator/ or for example a mathematical operator of formula/
Default Present: allanj, Laura, AWK, Glenda, Greg_Lowney, ScottM, JF, Katie_Haritos-Shea, MichaelC, KimD, jasonjgw, dboudreau, Lauriat, shwetank, Makoto, Melanie_Philipp, Detlev, Rachael, Wayne, MikeGower, marcjohlic, Bruce_Bailey, Jake, steverep, jon_avila, adam_solomon, Elledge, Mike_Pluke, kirkwood, Kathy, LisaSeeman, Michael, seeman, DAVID, Shawn, Joshue108, JohnRochford
Present: allanj Laura AWK Glenda Greg_Lowney ScottM JF Katie_Haritos-Shea MichaelC KimD jasonjgw dboudreau Lauriat shwetank Makoto Melanie_Philipp Detlev Rachael Wayne MikeGower marcjohlic Bruce_Bailey Jake steverep jon_avila adam_solomon Elledge Mike_Pluke kirkwood Kathy LisaSeeman Michael seeman DAVID Shawn Joshue108 JohnRochford
Found Scribe: Glenda
Found Scribe: glenda
Inferring ScribeNick: Glenda

WARNING: No "Topic:" lines found.

Found Date: 20 Apr 2017
Guessing minutes URL: http://www.w3.org/2017/04/20-ag-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.

WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report

[End of scribe.perl diagnostic output]