FHIR Chat · ADA/Screen-reader compliant IGs · ig publishing requirements

Stream: ig publishing requirements

Topic: ADA/Screen-reader compliant IGs

Elliot Silver (Sep 14 2021 at 17:13):

(Continued from #IG creation ADA/Screen-reader compliant IGs)

Does anyone have experience creating IGs that are compliant with jurisdictional requirements for accessibility? For example, so that screen readers can correctly navigate the IG, or there is sufficient colour contrast?

Grahame Grieve (Sep 14 2021 at 17:19):

with regard to colour contrast, this has come up before, and I don't use colour alone to convey meaning (at least, my policy is not to, and I've been relatively careful about that)

Grahame Grieve (Sep 14 2021 at 17:20):

browsers choose to highlight links in a different colour, but I don't think we interfere with that anywhere, because of accessibility

Jose Costa Teixeira (Sep 14 2021 at 17:21):

the template does have a variable for hyperlink color

Grahame Grieve (Sep 14 2021 at 17:21):

oh? I don't think it should - we should never say anything about hyperlink colour at all

Jose Costa Teixeira (Sep 14 2021 at 17:23):

I don't recall exactly what, but I think the project.css or the fhir.css were defining that, so I made it a variable

Grahame Grieve (Sep 14 2021 at 17:24):

we had some discussion before about navigability. I try to keep things direct and linear, for this reason, but we were going to hear back from some accessibility research that someone was doing, and I don't think we ever did

Jose Costa Teixeira (Sep 14 2021 at 17:24):

making it a variable doesn't mean it enables or disables changing color - it's all in css, so templates can change it anyway. Making it a variable just makes it less painful

John Moehrke (Sep 14 2021 at 17:27):

these are why I have only changed a very few colors, mostly to change header/footer background and thus the text that are used on those... Leaving all other colors as default.

Grahame Grieve (Sep 14 2021 at 17:28):

we don't fix hyperlink colors generally in fhir.css. We do it in specific circumstances e.g. in the xml examples

Elliot Silver (Sep 14 2021 at 17:28):

Not overriding the link colour could actually reduce navigability if the link colour is too similar to the document colour.

Grahame Grieve (Sep 14 2021 at 17:29):

and a in pre - fix to dark green. I'm not sure why we do that, and it doesn't seem like a good idea to me now, though I can't imagine who else would've done that

Grahame Grieve (Sep 14 2021 at 17:30):

if the link colour is too similar to the document colour

either the browser is 2wrong, or the document colour is set weirdly, which we never do. (except in the xml and json examples...)

Elliot Silver (Sep 14 2021 at 17:35):

Pointing wave.webaim.org at the IPS (https://wave.webaim.org/report#/http://hl7.org/fhir/uv/ips/), I get errors for empty buttons, missing alt text, low contrast, and missing H1 content. (Not picking on IPS--I think the issues detected are in the base template.)

Eric Haas (Sep 14 2021 at 17:42):

I used an app to score the contrast on the highlighting and I tweaked the colors to get close. forgotten all the details.

Grahame Grieve (Sep 14 2021 at 17:42):

missing H1 content? what's that about?

Grahame Grieve (Sep 14 2021 at 17:43):

where do we have empty buttons?

Grahame Grieve (Sep 14 2021 at 17:43):

low contrast - header or footer, for sure. I'm careful elsewhere

Grahame Grieve (Sep 14 2021 at 17:44):

missing alt text - hmm. that's kind of a problem. make a task to get me to be consistent there. Though hell, imagine listening to a reader reading a structure definition table.... what a nightmare. I don't know what else we could do about that

Elliot Silver (Sep 14 2021 at 17:45):

There is no requirement to be useful, just compliant. :laughing:

Elliot Silver (Sep 14 2021 at 17:47):

The missing H1 is that the highest level item on the page is a H2 header. There is no H1 header.
image.png

Grahame Grieve (Sep 14 2021 at 17:50):

I knew that. What I was really asking was, why would we care? And I am very resistant to changing that indeed.

Grahame Grieve (Sep 14 2021 at 17:51):

the way the header levels work is very deeply baked into the way we build content. Why would it matter whether a page has an H1?

Grahame Grieve (Sep 14 2021 at 17:51):

some pages only have a h3 or h4

Elliot Silver (Sep 14 2021 at 17:54):

I'm just beginning to look at this, so I don't yet know the implications in terms of accessibility of a page starting at H2 or 3 or 4. I just noticed that this is causing errors according to that website.

Yes, I imagine changing to generate starting at H1 would break a lot of things.

Elliot Silver (Sep 14 2021 at 17:59):

I think the "empty button" error is due to the navbar, but not sure.

Elliot Silver (Sep 14 2021 at 18:00):

<button data-target=".navbar-inverse-collapse" class="navbar-toggle" data-toggle="collapse" type="button">
              <span class="icon-bar"> </span>
              <span class="icon-bar"> </span>
              <span class="icon-bar"> </span>
</button>

Grahame Grieve (Sep 14 2021 at 18:04):

wtf is that?

John Moehrke (Sep 14 2021 at 18:06):

Elliot Silver said:

I'm just beginning to look at this, so I don't yet know the implications in terms of accessibility of a page starting at H2 or 3 or 4. I just noticed that this is causing errors according to that website.

Yes, I imagine changing to generate starting at H1 would break a lot of things.

some of us do use lots of H levels, so starting at 2 hurts.. but we deal with it. I can imagine how H1 is used by ADA readers to understand if they have covered everything worthy of covering.

Elliot Silver (Sep 14 2021 at 18:11):

Grahame Grieve said:

wtf is that?

Are you asking where I got that snippet, or what the heck that code is doing there? It's from http://hl7.org/fhir/uv/ips although I also see it at http://build.fhir.org/ig/HL7-Canada/ca-baseline, so I assume it's the base template.