W3C home > Mailing lists > Public > public-low-vision-a11y-tf@w3.org > November 2016

LVTF Meeting Minutes - Nov 3, 2016

From: Erich Manser <emanser@us.ibm.com>
Date: Thu, 3 Nov 2016 11:44:18 -0500
To: "Low Vision Task Force" <public-low-vision-a11y-tf@w3.org>
Message-Id: <OFDF454170.356E0BCE-ON85258060.005BCE47-85258060.005BF288@notes.na.collabserv.com>

Link to minutes with full text pasted below:

Low Vision Accessibility Task Force Teleconference
03 Nov 2016

See also: IRC log

      alastairc, Laura, Jim, Shawn, ErichM, Scott, JohnRochford, Glenda
         1.	Prioritization of what we can do in 4 meetings
         2.	Contrast: Informational Graphics
         3.	interactive contrast

      Summary of Action Items
      Summary of Resolutions

<ScottM> welcome back jim!

Happy to scribe today :)

<allanj> scribe: erich

JA: Looking ahead to the next 4 meetings for priorities. WCAG is looking to
get first public working draft out in Feb 2017
... Since we're all members of the WCAG wg, feel free to participate in
that also, to explain what we've done here, etc.
... With the Thanksgiving holiday, we've got this meeting plus 3 others.

Prioritization of what we can do in 4 meetings


JA: It seems our first 4 or 5 SC are the closest we have to complete
... All the others seem to be part of customization, could be addressed by
other work, taskforces
... Suggesting we focus on the first group, closest to being done, as

<allanj> +1

GS: Agrees with that

<JohnRochford> +1

LC: Also agrees


WD: Agrees provisionally, but Dec 1st deadline limits us\

GS: Let's get these solid and in by the deadline, and then when we have a
chance to breath move to the others

SH: Let's buckle down and do it rather than talk about it, get as much to
them as we can

WD: Element level customization needs attention

SH: We really must focus, next 2 or 3 meetings are critical

WD agrees

Contrast: Informational Graphics


JA: Did search on some info graphics that are up there, they are horrid
... If I have my infographic on a color wheel, there is no way I can make
it meet 4.5:1

GS: Pattern, pattern, pattern

JA: I see it, differentiated pattern is part of it, ok
... How do we feel about it?

AC: One question from TPAC still needs addressing
... Do we need some way of saying what the graphic must contrast with?

LC: I agree, Glenda did a great job in the other contrast one, Interactive
Elements, could we borrow from that?

<allanj> interactive elements contrast -

GS: I defined it as the visual presentation of important (non-text)
information... think that covers it, what do you guys think?


AC: Concerned less on icons, but more on larger info graphics. Realize the
patterns help, but SC text is saying it should have a contrast ratio

JA: Just added an infographic sample from Google
... It's not text, it's a graphic, so you can't get in to it

AC: There's a good example from the 1948, where does the orange and blue
need to contrast against the background, as well as each other

LC: That would be the same for a pie chart too

GS: Need to take a small step back, no need to handle in color contrast,
actually color alone causing the problem

SH: If I have pie chart and data table right next to it, it doesn't matter
if I can see the colors. No need to overcomplicate

AC: That would just mean we need some caveats, if it's being relied on and
there's no redundancy it should have sufficient contrast, or are you saying
it's not needed at all?

<Glenda> this is an interesting info graphic I was looking at last night

SH: What else do we need to say, do adjacent colors not need to have
sufficient contrast because we're already saying we're not relying on color

GS: Why don't we put that as an assumption, but for now think of less is

JA: This seems to fall in 1.4.5 in WCAG, Images of Text
... Wouldn't infographic example I added already fail?

GS: There are some fun exceptions
... Not trapping text inside a graphic is very important

JA: What is it in our SC that we need to say something must be done to fix,
and have we said that?

<allanj> rrsagent: make minutes

WD: Would like to add a note: There is not a lot of explicit reference in
the understanding document as to their application to low vision, things
about keyboard use and other things

<alastairc> How about: The visual presentation of graphics that *are relied
on to* convey important information have a contrast ratio of at least
4.5:1. (Using the 'important information' to define what needs the

WD: May need to look at the techniques of using color only, embedded
images, etc to see if there is an interpretation
... Need to use some asterisks along here

JA: Would like to focus on getting our language right in the SC

WD: Understood

JA: Does the language we currently have sufficiently convey there are
issues needing to be fixed?

GS: Proposes adding a sentence similar to:

<Glenda> Propose adding a similiar sentence like this to “Informational
Graphics”: The visual presentation of important (non-text) information in
an interactive image has a contrast ratio of at least 4.5:1.

<allanj> ACTION: jim with LVTF to review wcag understanding document and
techniques adding implications for Low Vision [recorded in

<trackbot> Created ACTION-85 - With lvtf to review wcag understanding
document and techniques adding implications for low vision [on Jim Allan -
due 2016-11-10].

LC: We would need to define, since pulling from COGA

<Glenda> Use COGA's important information definition:

<Glenda> important information

<Glenda> 1. information the user may need to complete any action or task
including an offline task.

<Glenda> 2. information the user may need to know related to safety, risks,
privacy, health or opportunities

AC: Is it the important information that needs to be contrasted with
adjacent items?

GS: I would say background, then that follows the pattern for tests


<Glenda> COGA definition on important information is located here:


<ScottM> +1

<ScottM> +q

<Glenda> +q

SM: Does not think asking people to check all possible color combinations
is practical. Would almost say it woudl be easier to indicate to use
information conveyed solely through color as the SC, and suggest removing
all color to test whether same information can still be conveyed

GS: This is an important gap which needs to be filled in 2.1
... There's something that's important to convey that is not text, but as
important as text
... We're not talking about differences between that pie slice and every
other pie slice, but can I see the image / important information

JA: There are different ways of seeing it, we have SC talking about 4.5:1,
but all of our techniques (GARBLED)

<ScottM> Anyone else hearing distorted audio?

<Zakim> allanj, you wanted to say focusing on contrast ratio, but
techniques list many other ways to fix. but sc says only 4.5:1

AC: Sounds like the hole we're trying to plug is discerning 'the thing'
from it's surroundings
... Status icons and warning images are more obvious use cases where we
could simplify
... Don't think we need to get in to slices of a pie chart, as that is
covered by the color alone aspect

<Zakim> AWK, you wanted to say that what I hear Glenda talking about is not
relying on color alone

<Glenda> I agree with what alastair just said. This is about
discerning/seeing an item compared to the immediate surroundings.

AWK: We want to make sure things like status indicator warnings, if text,
would be 4.5:1

SC: Covers status and focus in SC I am working on

GS: Covers status and focus in SC I am working on

<alastairc> So add to the SC text something like: "The visual presentation
of icons or graphics that convey important information have a contrast
ratio of at least 4.5:1 against the background".

<AWK> +1 AC's suggestion

WD: Need to see the line indicating the changes, that's what needs to be

<AWK> But it doesn't cover the use case of colored elements within a

GS: Proposes that we add 'with it's immediate surroundings' to the end of
what's currently written

AWK: If I make a line graph with two lines that follow a grid, 1 dashed
line and 1 solid line, that might satisfy what Wayne is saying

<Glenda> Here is a timely example:

<allanj> provided the lines are contrasting enough with the background

<Glenda> scroll to bottom to see line graph

AWK: We're getting in to real specific language that risks not having broad

JA: Glenda example just posted fails SC as Alastair has posted

<Glenda> The yellow to white is 1.8 to 1

JA: The black text passes, for state names, but the white outlines, not so

AWK: In state map situation, are we trying to differentiate between these
different colors as well?
... You can have one color that's touching many others

GS: It's okay, in those outlines, it's immediate surroundings
... I was actually looking at the line map at the very bottom, and just
looking at the yellow, blue, red, and asking if they can all be 3:1 since
they're all big things

AWK: On that same chart, do the grid lines need to be contrasted also?

GS: Is it important?

AWK: Argument could be made that it is

<allanj> "The visual presentation of icons or graphics that convey
important information have a contrast ratio of at least 4.5:1 against the

JA: So going back to AC's text, have we conveyed what we're trying to say
... Does the language address all the things we've talked about regarding
issues we have with the particular image

AWK: no

<Glenda> Propose changing language to 'The visual presentation of important
information in icons or graphics has a contrast ratio of at least 4.5:1
against the immediate surrounding background.”

AC: We would almost have to define which elements are important to
... Need to come up with some frame around discernibility to understanding

JA: What sort of language do we need to work in to Glenda's proposed
language to address discernibility to understanding?

<Glenda> Use COGA's important information definition:

<Glenda> important information

<Glenda> 1. information the user may need to complete any action or task
including an offline task.

<JohnRochford> +1 to using COGA definition of important information

JA: And we need to differentiate this from just the basic color contrast SC
... Trying to see how we separate color alone to other information

WD: Why does it fail color alone?

JA: It's the yellow line, I can't see it

GS: That's discernment

AC: It's the definition of graphics that's the problem here
... when we get in to more complex graphics, they need to be discernible
against each other
... If we consider blue line, red line, yellow line as graphics, one surely

<Glenda> Suggest we handle it with examples (like this line chart and state
chart for the election)

AC: if it's important to understanding, it should have sufficient contrast
against it's background

GS: I propose that SC language suggested is within reason, and be
supplemented with additional examples

<allanj> AC isn't this covered under WCAG 1.4.3 Contrast (minimum)

<JohnRochford> Gotta go, folks.

WD: Informational text has a foreground and a background
... IG is generally trying to depict something, so in the foreground you
can tell what it's depicting against clearly a background

<AWK> what would we say about: https://finviz.com/map.ashx

<Glenda> each informational element in an information graphic must have a
color contrast of 4.5 to 1 with it’s immediate surrounding.

AC: In complex graphics, need to say each complex graphic is made up of
parts of a graphic, and if important to understanding, needs to have its

GS: Inside an icon, may have black icon with white smiley face, and that
face is immediately surrounding the black, so it works and is simple

WD: What if your background color is dark charcoal

JA: You could see the white smiley face, but would you know that anything
is surrounding it
... Are we any closer on this one?

WD: If we combine what AC did with GS clarification

<allanj> proposal: each informational element in an information graphic
must have a color contrast of 4.5 to 1 with it’s immediate surrounding.

AC: This could use more description, I will tackle for next week

<Glenda> then we will need to add 3 to 1…can we look at my stuff

JA: We keep saying 4.5:1, but 3:1 for larger things



<shawn> +1 for sticking with 4.5:1 and 3:1 (and 7:1) from WCAG 2.0

GS: Using 4.5:1 as default, and less comfortable but left it also for
disabled elements

<laura> 1 for sticking with 4.5:1 and 3:1 (and 7:1) from WCAG 2.0

GS: What is the normal pixel size at 14 pt, I did not measure for that

JA: Alastair will work more on the description. Seems we have loose
agreement on proposed language, though we did not have a vote
... Thinks description will tease out the size aspects

AWK: I have concerns, very complicated and talking about many complicated
things. Worried what we come up with may fail dramatically
... We may wind up relegating to AAA because it's not specific enough about
what we need to do

GS: Thinking a lot about it, and hoping your concerns is a normal concern
at the beginning, but that you would start to become more comfortable with
... Comes back to questioning, why did you put that in the graphic, did you
want somebody to see it or not?

<Glenda> AWK, would you be willing to go with 3 to 1? and you just put
borders around it…

AWK: Concerned about instances where there are many adjacent colors, and if
one of those colors does not comply, the whole thing fails the SC

GS: I am not a designer, but I wonder if there are designers in the group

AC: I do work a lot with designers, and feel it's possible to do this, if
we narrow and use patterns, labels

<Wayne> q

GS: It's being able to see the line, or to see the label. It's not getting
in to interpreting the legend, but just merely it could enter my brain

AWK: agreed

GS: I would be willing to start with 3:1 if concern is high, or we could
start with 4.5:1 and back down to 3:1 if needed

AC: If anyone has come across examples (good or bad) please send to me

WD: THink the problem with IG from a visual point of view, they are
important but cannot be changed programmatically
... there's no way with style that you can fix that
... I like AC's idea of narrowing it, I think we should narrow this to what
we can see right now, with the risk or certainty we'll leave something out

GS: sending data visualizations example with Alastair

<Glenda> Data Visualizations to consider:
http://www.informationisbeautiful.net David McCandless

WD: we could look at Edward Tufte (sp?) examples also

interactive contrast

GS: Do we want to start with 4.5:1 or simplify and call everything 3:1 to
start (would make it much easier, but not sure it's enough)

<alastairc> AC: 3:1 might be ok with a minimum pixel width of 3px (maybe
2?), but with 1px width then 4.5:1 would be needed.

<shawn> [ I wonder if information not conveyed by color alone covers the
deisabled issue]

<shawn> [ also with high contrast, it might be too hard to have sufficient
change in contrast to be able to tell the difference between enabled and
disabled ]

<allanj> Erich: +1 to 3:1 for disabled elements

<allanj> wd: +1 also

AWK: Do we know default contrast ratio of the browser for disabled

GS: no

AWK: Are we comfortable saying you fail if you use the default?

GS: This would be a change from 2.0, since 2.0 lets you get away with the
... If we're trying to move a11y forward and see the focus, then it's not
good enough, can we bring this forward and have the possibility of
discussing with the browser vendors to make better

<Zakim> AWK, you wanted to ask what the contrast ratio is for default
browser controls when disabled

<laura> +1 to goodwitch

<allanj> jim will create a chart of contrast for form elements and focus

<alastairc> default outlines on inputs in Chrome: 2.14:1 for regular,
1.41:1 for disabled! (1px wide)

<shawn> +1 for developing additional material, e.g., adding to existing
Understanding docs -- and maybe more supporting material for the new SCs

<alastairc> Quick test:

<allanj> thanks!!!

<alastairc> Glenda - can you update the wiki with your contrast SC text?
Not sure I saw the right one go past...

rrsagent: make minutes
                    Erich Manser                                                       
                    IBM Research                                                       
                    MA / tel:                                                          
                    Search for                                                         

You don't need eyesight to have vision.

(image/jpeg attachment: 0F334665.jpg)

(image/jpeg attachment: 0F925230.jpg)

(image/gif attachment: ecblank.gif)

(image/gif attachment: 0F527542.gif)

Received on Thursday, 3 November 2016 16:46:00 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:23:23 UTC