button table { margin: 0; } .fa-arrow-left::before {content:"←"} Bad code <button type="button" onclick="window.open('https://example.com/other-page')"> <table> <tbody> <tr> <td> <i class="fa fa-arrow-left" aria-hidden="true"></i> </td> <td align="left">Back</td> </tr> </tbody> </table> </button> Back Issues and how to fix them The HTML isn't valid because the table element isn't allowed as a child of the button element. Based on the content categories listed on MDN, the table element qualifies as flow content and can only be placed inside an element that expects flow content. The button element, on the other hand, expects phrasing content, not flow content. The semantic information of the table doesn't cause any issues because the button's content is presentational See Buttons Role on W3C, Presentational Children. Keep it simple and structure the button element using CSS, such as Flexbox. Instead of using tables to structure the button element, use phrasing content elements, such as span, img, or SVG. Good code <button type="button" onclick="alert('Hello World!')"> Back </button> Resources <button> on MDN <table> on MDN Content categories on MDN Buttons Role on W3C <button> on W3Schools <table> on W3Schools Presentational Children

by Rian Rietveld It's a common pattern to use a logo in the header as a link to the homepage. Fun fact: the alt text of the image inside a link, will be added to the link text. The problem with linking a logo is that it serves 2 purposes: a logo, that tells you which site you are visiting; a link, that leads to the homepage. So add both that info in the alt text. Explain what's on the image and where the link leads to: "Site name logo, to the homepage". Note: Start the alt text with the visible text, then the link will be easier to target for people using voice recognition software. Let's take as example the logo of my home town Leidschendam-Voorburg. In code (simplified): <a href="/"> <img alt="Leidschendam-Voorburg logo, to the homepage" src="logo-lv.svg" /> </a> Generated HTML: VoiceOver in Safari will announce this as: All info is there. There is an advantage of using the alt text instead of an aria-label solution. When the connection is slow, the alt text will show up before the image does and already informs all users of the site name and link destination. It's quite robust. Discussion Do I need the word logo? Isn't that redundant, like the word image or picture? Yes, in this case the fact that this is a logo tells important info. Not only for blind users, but for all screenreader users. If you are visually impaired but not totally blind, you may see the image vaguely. With the addition of the word logo, you also know what the image is. The site name is required, but the logo gives extra context. Isn't this a violation of 2.5.3 Label in Name? No, the visible name is in the alt text, which becomes the accessible name of the link. Adding extra context to the link is common practice. Compare this with the notorious "Read more" link in a card. A common pattern is to use a sr-only CSS class (or equivalent) with additional text: <a href="url">Read more<span class="sr-only"> about cute kitten</span></a> . You see Read more. You hear Read more about cute kitten. In our case: You see the site name and logo; the visual position tells the link destination. You hear the site name, the fact it's a logo and the link destination. The whole visible text should be in the accessible name, but extra information, like the meaning of the image and link destination can be added for context. That way the purpose of the link is explained. Why didn't you use an aria-label? Of course there are other and also valid solutions. By using plain simple HTML I want to honor the first rule of ARIA: If you can use a native HTML element or attribute with the semantics and behavior you require already built in, instead of re-purposing an element and adding an ARIA role, state or property to make it accessible, then do so. Happy Holidays!

Bad code <button type="button" onclick="window.open('https://example.com/other-page')">Link target description</button> Issues and how to fix them A button opening a link will be unexpected behavior for screen reader users. No matter how it is styled. Links disguised as buttons won’t show up in the link list of a site in assistive technologies. Use links for navigation to other pages or sections, and buttons for actions performed on the current page or within the application. Good code <a href="https://example.com/other-page">Link target description</a> Resources ` on MDN](https://developer.mozilla.org/en-US/docs/Web/HTML/Element/a) - [`` on MDN](https://developer.mozilla.org/en-US/docs/Web/HTML/Element/button) - [Buttons vs. Links by Eric Eggert](https://yatil.net/blog/buttons-vs-links)

The good intentions were there but in the HTML and Accessibility world, less is sometimes more. Bad code <label for="textinput">First name</label> <input type="text" id="textinput" aria-label="First name" placeholder="First name" title="First name"> Issues and how to fix them The aria-label, placeholder, and title attributes all provide the same informaton ("First name"), leading to the screenreader reading the same text multiple times. The aria-label is unnecessary since the input is already correctly labeled by the <label> element. <label> elements should be preferred to aria-label since they are also visible to sighted users. The title attribute is not needed here, as the label and placeholder already convey the necessary information. Using title in this context adds unnecessary complexity. The placeholder text should provide a hint / example to the user what kind of input is expected, it should not act as a label or contain the same content. Good code <label for="textinput">First name</label> <input type="text" id="textinput"> Resources <label> on MDN aria-label on MDN placeholder on MDN title on MDN

by Vasilis van Gemert Recently when I gave a coding assignment — an art directed web page about a font — a student asked: does it have to be semantic and shit? The whole class looked up, curious about the answer — please let it be no! I answered that no, it doesn’t have to be semantic and shit, but it does have to be well designed and the user experience should be well considered. Relieved, all of my students agreed. They do care about a good user experience. The joke here is, of course, that a well considered user experience starts with well considered HTML. Somehow my students are allergic to semantics and shit. And they’re not alone. If you look at 99% of all websites in the wild, everybody who worked on them seems to be allergic to semantics and shit. On most websites heading levels are just random numbers, loosely based on font-size. Form fields have no labels. Links and buttons are divs. It’s really pretty bad. So it’s not just my students, the whole industry doesn’t understand semantics and shit. Recently I decided to stop using the word semantics. Instead I talk about the UX of HTML. And all of a sudden my students are not allergic to HTML anymore but really interested. Instead of explaining the meaning of a certain element, I show them what it does. So we look at what happens when you add a label to an input: The input and the label now form a pair. You can now click on the label to interact with a checkbox. The label will be read out loud when you focus on an input with a screenreader. When you hover over a label, the hover state of the connected input is shown. My students love stuff like that. They care about UX. I show them that a span with an onclick-event might seem to be behaving like a link, but that there are many layers of UX missing when you look a bit closer: right-click on this span, and a generic context menu opens up. When on the other hand you right-click on a proper link like this one a specialised context menu opens, with all kinds of options that are specific to a link. And I show them that proper links show up when you ask your screen reader to list all the links on a page, yet spans with an onclick-event don’t. Moreover, a span doesn’t receive focus when you tab to it. And so on. My students see this and they get it. And they love the fact that by being lazy they get much more result. So when I teach about HTML I always start with the elements that are obviously interactive. I show them the multitude of UX layers of a link, I show them the layers and layers of UX that are added to a well considered form. I show them what happens on a phone when you use an input with a default text type instead of the proper type of email. They get this, and they want to know more. So I show them more. I show them the required attribute which makes it possible to validate not only form fields, but also fieldsets and forms. They love this stuff. Then I go on and show the details element with a summary. Whoah! No JavaScript! These interactive things are the most important parts of HTML. These are the things that truly break or make a site. If you don’t use these elements properly, many people are actively excluded and the UX degrades in many ways. The interactive elements are what makes the web the web. The old focus on semantics Every time I come home from a web conference I hang the lanyard with the badge on a doorknob. There are at least 35 lanyards there. The oldest is more than fifteen years old, the newest just a few months. On most of these conferences there was someone telling us about the importance of using proper semantic HTML. And they all talked about the importance of heading levels. So for at least 15 years we’ve been telling the web design and web development community that heading levels are very important. Yet after all these years, and after all these conferences there are almost no websites that do it right. Later, when HTML5 was introduced the talks about semantics were still mostly about creating a proper document outline, but this time with sectioning elements in combination with headings. Unfortunately this idea was never properly implemented by browsers. And it turned out that these sectioning elements are very hard to understand. For years I’ve tried to explain the difference between a section and an article, and almost nobody gets it. And of course they don’t, because it’s very, very complicated theory. And above all: There’s no clear UX feature to point at. The user experience doesn’t change dramatically if you use a section instead of an article. It’s mostly theoretical. So nowadays, next to divitis, we also have sectionitis. Understanding how heading levels work is hard. And understanding how sectioning elements work is really hard. And explaining these things is even harder. What’s the use for theoretical purity if the end result is the same? Yes, I know that some sectioning elements actually have some UX attached to it. But not that much UX if you compare it to the real interactive elements. Not getting your heading levels right is not at all as destructive as using divs instead of links. Now, I’m not saying that we should stop teaching people about heading levels and sections. We shouldn’t. Heading levels and sections do things as well. But we should think about when we teach these more complicated and subtle parts of HTML. First we need to get people exited about HTML by showing all the free yet complex layers of UX you get when you use the interactive elements properly. And then, when they do understand the interactive elements, when they’re really excited and they ask for more, show them the more obscure UX patterns. You need a good idea of what UX is before you can understand things like the option to nagivate through the headings on the page with a screen reader. Without an idea of what UX means you cannot understand what landmarks do. First start with the obvious, then show the details. I am very much looking forward to 15 more years of web conferences and publications. I look forward to seeing inspiring talks about the UX of HTML. Talks about the incredible radio button and its wonderful indeterminate state! Talks about validating forms in a friendly way with just HTML and some clever CSS. Talks about the different context menus that appear on different elements. In other words, talks about what HTML does, and much less about what it means in theory. Let’s talk about user experience, and let’s stop talking about semantics and shit.