I've got a list of buttons that filter a catalog in a SPA app. The structure looks like this:
<div>
<button>filter 1</button>
<button>filter 2</button>
<button>filter 3</button>
</div>
<div>
{results}
</div>
Clicking each of button causes new results to be fetched from a server and then updated in the area {results}.
We've got an accessibility expert that is insisting that we change the above structure to be a tablist, but I don't see how it would work, as from my understanding, a tablist needs tabs and a corresponding tabpanel. The buttons could be the tabs, but what of the tabpanel? I don't have the data pre-existing, it must be fetched when the buttons are clicked.
How would I update the above to make it comply with accessibility?
It kind of feels like a tablist because clicking on the buttons essentially changes your view but there is not an accessibility reason that you have to use the tablist design pattern.
I agree that you'd sort of be "forcing" it into the pattern since you only have one panel that updates rather than three separate panels.
It's perfectly fine to have a group of 3 buttons all update the same container and not use a tablist.
Is it obvious to the non-sighted user that they have 3 buttons to choose from? If not, it might be nice to have them contained in a list. You can use a real <ul> and turn off the bullet point styling or use role="list" and role="listitem".
<div role="list">
<span role="listitem">
<button>filter 1</button>
</span>
<span role="listitem">
<button>filter 2</button>
</span>
<span role="listitem">
<button>filter 3</button>
</span>
</div>
I used <span> containers for the buttons to keep the inline nature of your original example.
Next, is it obvious to the non-sighted user that the page has updated when the button was selected? It might be as simple as saying "selecting a button will update the results" or it might be obvious from the button labels. Depending on your situation, aria-live might be needed. But that's going beyond the extend of your original question.
Related
How to select Actions Button followed by the Div class containing known Text ( for example, card_header-title"Addresses" in this case) in Robot Test Framework?
The page contains several span table sections and each of them has its own actions and show-history buttons. To select the specific Actions button, I could use its xpath, but I am trying to access all sections in a for loop and the xpath of actions button in one section changes from the other, so hard coding is not an option for me. Would someone please help.
<div class="attribute-group-header card__header">
<h3 class="attribute-group-title card__header-title">Addresses</h3>
<div class="floatright">
<input type="button" class="action small btn" value="Actions">
<input type="button" class="showHistory action small btn" value="ShowHistory">
</div>
</div>
I know you say you don't want to use Xpaths but maybe one of these examples could help. I don't see any other way of achieving what you're asking for other than having id's supplied on the buttons.
You could use an xpath locator that first finds the text of the "attribute-group-title card__header-title" element and then selects the following sibling div, followed by the input:
//*[contains(text(),'Addresses')]/following-sibling::div[1]//input[#value='Actions']
So, I have the IssuesList component, which is the list of issues that I get using ajax and github api, and DevStatus component, which sort of wraps the list up and contains all the logic, triggers state changes by two radiobuttons and so on.
My problem: When I click on one of the radiobuttons, the DevStatus component won't change state if the click was on the text inside the radiobutton. And when I click on the corners of the radiobuttons, the blue areas without text, the state changes perfectly.
Here's the structure of the radiobuttons:
<div className="btn-group" data-toggle="buttons">
<label className="btn btn-primary active"
onClick={this.onChangeRadioButton.bind(this)}
id={this.CLOSED_ISSUE_ID}>
<input type="radio" name="options"
autoComplete="off"
id={this.CLOSED_ISSUE_INPT_ID}
onChange={this.onInputChange.bind(this)} /> Closed Issues
</label>
<label className="btn btn-primary"
onClick={this.onChangeRadioButton.bind(this)}
id={this.OPEN_ISSUE_ID}>
<input type="radio" name="options"
autoComplete="off"
id={this.OPENED_ISSUE_INPT_ID}
onChange={this.onInputChange.bind(this)} /> Open Issues
</label>
</div>
Here's the codepen with the code and here's the full page view so you could better see and understand what I'm talking about.
Please, open the full page view and try to click on parts of the button that contain text and on ones that don't and you'll notice that as long as you click on parts without text - the state changes and if you click on text itself - the state doesn't change at all.
Could you please help me with that problem?
PS: removing onChange from the input element is not the solution.
Update 1
If you go to DevTools and inspect the radiobutton element, you'll see that inside the label tag there're input and weird span elements. The span element is not in the code I wrote, did React automatically add that? For some reason, the onClick event listener is not applied to those input and span elements.
Update 2
I've tried to add click event listener to the radiobutton in the console of dev tools and tried to figure out the target of the clicked element. When I click on the text - it is the span element and when I click on place without text - it is the label element and that's why the click event is not working.
Can my problem be solved using dangerouslySetInnerHTML, so that it won't create the unnecessary span?
Could you tell me please how to solve that?
React is creating a span because your text is not in any div. Also it would create a span if there was any white space (but in your case this is because there is no div around your text).
But the real problem here is the way you check your event. You need to check e.currentTarget instead of e.target
Then no need to use the ugly dangerouslysetinnerhtml!
React appeared to sometimes be adding span tags around text, no matter if there are the free white-spaces or not. The spans didn't allow the onClick event to fire when they were clicked on.
So, to force React not to render the spans, the dangerouslySetInnerHTML may be used:
noSpanRender(text) {
return { __html: `<input type='radio' name='options' autoComplete='off'/>${text}` };
}
render() {
return (
<div className="dev-status-page col-centered">
<div className="graphs">
<h1 className="text-center page-header">
Our Recent Closed and Opened Issues from GitHub
</h1>
</div>
<div className="issues col-centered">
<div className="btn-group" data-toggle="buttons">
<label className="btn btn-primary active"
onClick={this.onChangeRadioButton.bind(this)}
id={this.CLOSED_ISSUE_ID}
dangerouslySetInnerHTML={this.noSpanRender('Closed Issues')} />
<label className="btn btn-primary"
onClick={this.onChangeRadioButton.bind(this)}
id={this.OPEN_ISSUE_ID}
dangerouslySetInnerHTML={this.noSpanRender('Open Issues')} />
</div>
<IssuesList issues={this.state.issues} />
</div>
</div>
)
}
It was vital to avoid those span elements inside the input tag, so using dangerouslySetInnerHTML finally helped.
I'm wondering about screen reader accessibility using Twitter Bootstrap framework and FontAwesome icon fonts.
I'm looking at 2 different icon situations:
1) The icon has helper text that a screen reader will pick up:
<span class="fa fa-pencil"></span> Edit
2) And a standalone icon without any helper text:
<span class="fa fa-pencil"></span>
Ideally, in both situations, a screen reader will announce that the element is an "Edit" button.
Per FontAwesome's site:
Font Awesome won't trip up screen readers, unlike other icon fonts.
I don't see any speech css tags related to FontAwesome or Bootstrap and not really clear to me how a screen reader will react to each of these situations.
I'm also aware of aria-hidden and Bootstrap's .sr-only and there has to be an ideal way to handle both situations.
Edit: added title="Edit to example 2.
What advantage does using aria-label="Edit" have over the standard title="Edit"?
Edit 2: I came across this article that explains pros and cons of different use implementations.
First of all, you should probably use <button> instead of <a href="#">. Empty links can be confusing for screen readers, but a button is a button. In short, links take you places, buttons perform actions. (http://www.karlgroves.com/2013/05/14/links-are-not-buttons-neither-are-divs-and-spans/; https://ux.stackexchange.com/questions/5493/what-are-the-differences-between-buttons-and-links).
I would go with a variation of your first code sample, and utilize Bootstraps .sr-only class. If we update your code with button and add in the class, we have:
<button type="button" class="btn btn-default"><span class="fa fa-pencil"></span> <span class="sr-only">Edit</span></button>
We now have a more semantically correct button element; sighted users see the edit pencil icon; and screen reader users will hear "Edit". Everyone wins.
(Note, the button code is straight from Bootstraps CSS Buttons section.)
From my understanding I think it may be useful to also add in:
aria-hidden="true"
to the span class that holds the pencil icon. This will prevent the screen reader from trying to read this element.
<span class="fa fa-pencil" aria-hidden="true"></span>
As part of making our site WCAG compliance we are adding play/pause buttons for carousel.Here the screen reader is reading in different manner
<div id="imageCarouselPlayBtn" class="cblt-button imageCarouselPlayBtn" tabindex ="0">
►
</div>
<div id="imageCarouselPauseBtn" class="cblt-button imageCarouselPauseBtn" tabindex ="0">
‖
</div>`
in this case screen reader is reading pause button as double vertical line and group for play button.
if we use sprite image instead it is reading as group for both play/pause buttons.
Is there any good solution for this problem instead of using image tags for both ?
You could use WAI-ARIA’s aria-label attribute:
It provides the user with a recognizable name of the object.
Also: Is there a reason why you don’t use button/input instead of div for the play/pause buttons? With buttons, your markup could look like:
<button type="button" aria-label="Play">►</button>
<button type="button" aria-label="Pause">‖</button>
I'm using jQuery Mobile (jQM) and Knockout.js (ko) to develop an application. In this application, I need to generate a variable number of buttons that are defined by a constantly updating web service.
So, in my markup, I have:
<div id="answerPage-buttons" data-bind="foreach: buttonsLabels">
<button data-role="button" data-inline="true" data-theme="b" data-bind="text: text, click: $root.submitAnswer" />
</div>
buttonLabels is a list of short strings returned from the web service. It's defined as:
self.buttonLabels = ko.observableArray();
This all works fine when the buttons are not "jQM styled". However, when I style them using:
$("#answerPage-buttons").trigger("create");
problems arise during the update.
The issue seems to be that jQM wraps the buttons in a div (with a sibling span) to make them all nice and mobile looking. However, when the ko applies the updates via the bindings, it only removes the tags, leaving the surrounding stuff, and adds new button tags - which are then also styled by the jQM trigger call.
So, I end up with an ever-growing list of buttons - with only the last set being operational (as the previous ones are gutted by the removal of their button element, but all the styling remains).
I've managed to address this, I think, by placing the following call immediately after the observable is updated:
$("#answerPage-buttons div.ui-btn").remove();
However, my feeling is that there's probably a better approach. Is there?
I found a solution.
If I surround the buttons with a div, it seems to work - e.g.
<div id="answerPage-buttons" data-bind="foreach: buttonsLabels">
<div>
<button data-role="button" data-inline="true" data-theme="b" data-bind="text: text, click: $root.submitAnswer" />
</div>
</div>
I'm guessing this is because the markup added by jQM remains "inside" the markup replicated by ko. Without the div, jQM wraps the button tag, which was the immediate child of the tag that contains the ko foreach binding.