1. <dl> has no corresponding (viz. implicit) role, but can be given the role group, list, none or presentation <https://w3c.github.io/html-aria/#el-dl>.
Wow I have never noticed that, thanks for the heads up! Out of curiosity, would you put `role="listitem"` on `<div>` elements that wrap the `<dt>` && `<dd>` elements? It looks like `role="listitem"` is allowed on the `<dt>` element, but that doesn't feel like it would be accurate in the case where multiple `<dt>` elements are grouped together, and I'm not sure if that would mess with how the element is interpreted natively as as term.
I pulled up some of those spec's examples in Chrome and viewed the accessibility tree, and weirdly it assigns <dt> an implicit role of `term`, even though the spec says that <dt> should be `listitem` and <dfn> should be `term`. I didn't check FF or Safari but I would not be surprised if they did something different.
Given the spotty screen reader support for dt/dd to begin with, I'd steer clear of using multiple `<dt>` s if a11y is top priority, and just modify your content to have one dt per dd. Your single dt text could be "X or Y", repeat the definition twice, etc.
If you're familiar with Jakob's Law, it applies to a11y too, so stick with what users might have seen on other sites.
This is going to be unpopular here, but life became easier when I quit trying to write semantic HTML. It’s just poorly designed, I’m sorry. Every time I’ve reached for a <dl> I’ve eventually regretted it because I wanted multiple levels of wrappers, or a divider between sections, or an icon, or a heading spanning multiple key-value pairs, etc. They make this stuff with some flexibility but nowhere near enough to actually cover the generalized concept it purports to. I still use the corresponding elements when there are observable benefits, of course, like <button>, <h1>, etc, but when all it’s going to do is not quite fit the data model and force me to override everything, it’s just not a practical choice.
It shouldn’t be so controversial to say that if 99% of usage routes around your API, it’s probably the API’s fault.
As a person who daily drives a screen reader, I so agree with this.
We'd all be better off if the W3 dispensed with all that ideological semantic purity BS and started doing more realpolitik. Think not about whether your API is semantically pure, but about what developers want to do, what hacks they'll use to achieve their goals despite your objections, and how to enable doing those things in a way that is maximally beneficial to everyone involved.
ARIA live regions are the perfect example. What developers actually want is `document.speakText`. What developers actually have is a weird API that announces text on the page as it changes. They have to bridge from one to the other, which is difficult and hacky, even when implemented well. But hey, at least that live region approach is semantically pure HTML...
I don't want that. I don't want to care about screen readers (unfortunately I have to). I want a system where I can pick well-defined rules and then css can style it, screen readers will understand it, automations can parse it, keyboard navigation is free.
Obviously thats not what we got, but I feel like the set of established UI patterns is manageable enough that it could be built.
A great example is the new <select> styling that developers styled in all kinds of creative ways. Now give me that for comboboxes, trees, data-grids etc...
Sounds like it's CSS' fault then. I think that just like they introduced `display:contents` to remove wrappers, they should also introduce a way to group elements as if they had a common ancestor.
With CSS Grid math you can fake it, at least. If your DL is `display: grid;` and if you have a few extra DIVs lying around at the bottom of the DL to be borders around combined cells you just have to math which rows/columns you want to draw a border around and make the div fit that combined shape.
Good idea. Together with ::after / ::before and content: (which can insert text into the website) it might then be possible to create a website without any HTML, only CSS.
I feel similarly about HTTP. The protocol maps really well to resource stores (stuff like S3). GET, PUT, DELETE all make sense. HTTP status codes are built exactly for this use case as well. But as web developers we're mostly not developing resource stores. Those are highly generic and can be built once and used by millions of apps. Most of the time someone is writing code that interacts with HTTP they are performing RPC. You can go for GraphQL, gRPC, or many other RPC systems that just shirk the whole thing. They make everything use a POST to a single endpoint and add another layer of abstraction so you don't need to return a 4XX/5XX error for some highly application-specific situation.
It's clear the RFC writers got a little out of hand. "402 Payment Required", "407 Proxy Authentication Required", "508 Loop Detected" look to me like attempts to work in functionality specific to certain types of apps or deployments. Why do these RFC authors get their specific needs implemented into the bedrock of the web and then expect me to find where my needs happen to overlap and then tuck every aspect specific to my app into "400 Bad Request" or "500 Internal Server Error"? Every time I see a web app actually utilize more than the bare minimum of HTTP status codes I roll my eyes. Put that shit into the application layer. The protocol wasn't made for you. It was made for LAMP-stack apps serving mostly static assets.
List history (listory?) lesson, kids: As the link below to a 1985 IBM mainframe DCF/GML manual shows, DL-DT-DD have been a thing since before the web. In addition to Definition lists (DL), the 40+ year-old documentation describes Glossary lists (GL), Ordered lists (OL), Unordered lists (UL), and Simple lists (SL).
ibm :: 370 :: DCF :: SH35-0050-2 Document Composition Facility Generalized Markup Language Implementation Guide Rel 3 Mar85
I believe IBM Generalized Markup Language evolved into SGML (Standard Generalized Markup Language) which was used heavily at CERN when Tim Berners-Lee was working on HTML, which is heavily derived from it. I find this interesting about HTML: some form of the markup language had been floating around for decades until Berners-Lee added hyperlinks to it.
GML dates to 1969, SGML from the 1970s. Internally we used something called BookMaster which kind of? looks like a precursor to HTML (you had :p. instead of <p>, :li. instead of <li>). There was an effort circa 1990-1991 (as TBL was developing HTML and HTTP) called HyTime which was an SGML application focused on hypermedia. HTML killed that fairly quickly.
> Prior to HTML5, this was called a definition list. This is because the <dl> was originally only intended to represent glossaries of terms and their definitions.
Eh, it's fine, elements should be defined for what they mean, not what they look like. The explanation and distinctions made between <b> and other elements (<i>, <em>, <strong>) make sense.
The suggested (not obligatory) user agent styling for <b> is `font-weight: bolder` an agent or authors could use lots of different things to bring attention to what the element contains and treat it differently from <strong>.
Great post. Very minor nitpick: ‘The small element must not be used for subheadings; for that purpose, use the hgroup element.’
‘The small element represents side comments such as small print. Small print typically features disclaimers, caveats, legal restrictions, or copyrights. Small print is also sometimes used for attribution, or for satisfying licensing requirements.’
I love DL. I think tables, at least in the past, were misused as DLs even more in the past and the inconvenience of the table markup is even worse than a bunch of divs.
As someone who had written lots of XHTML in the past, not having closing tags makes my eyes twitch like Scrat in Ice Age. I even occasionally write `<br>` like `<br/>` out of habit.
Fair point, though /DT and /DD are also optional just like /TH, /TD and /TR are. So in effect, def…scription list could structurally save you one TR for each entry and two "BLE"s:
As others have noted only the div is allowed. This isn't a unique situation either, the HTML spec despite being lenient in syntax is quite restrictive in behavior. It's unfortunate that XHTML (and XML parsing) didn't become the default as it's the opposite, more restrictive syntax, but lenient behavior.
For example in XHTML you can use custom elements as table rows or cells (provided you give them the correct role and CSS display property). This is because XHTML does not modify the tree during parsing, unlike HTML which will hoist out custom element children of the table to the table's parent.
They're not as you can see here[1] and here[2]. They both inherit from the HTMLElement interface but div is considered an HTMLDivElement which makes it distinct from a custom element.
Exactly! It cracks me up when people name-check "semantic elements" when it doesn't actually mean anything in that context. Accessibility software doesn't understand the semantics of your custom elements, so there is no benefit in that situation whatsoever. Maybe it's easier for you to read and edit in the future, but that's it.
Somehow, people got convinced that <div> elements are evil and should never be used no matter what. Yes, you should use a more semantic element when it makes sense, but try to remember what that phrase actually means.
Aside from the accessibility benefit mentioned in the article, I would think that using semantic HTML like this would also make AI tools more effective at understanding code
I've used this a good amount of times, when I coded in front end projects. The first time gave me that satisfying feeling of using the right tool for the job, like completing a puzzle of HTML semantics. I remember JAWS not announcing it correctly in 2018, not sure if it's better now.
When I checked in 2024 or 2025, Windows Narrator announced it differently in Chrome, Firefox, Edge (Chromium mode) and Edge (IE mode), and none of them worked how I would expect them to. Adrian Roselli's verdict (https://adrianroselli.com/2025/01/updated-brief-note-on-desc...):
> Description list support continues to be generally good (with VoiceOver still the outlier), even if you may not like how it is supported.
You shouldn't try to fix this kind of thing by mangling the HTML, since (1) users tend to be used to their screen reader's quirks, and (2) in situations like these, making it juuuust right in one screen reader is likely to make it incomprehensible in another. But it is important to be aware of these quirks, so you don't accidentally design an interface that relies on less-quirky behaviour.
> Admittedly, however, support for the <dl> element is not yet universal.
Wait what? <DL> has been in HTML since.. the first draft in 1993!
I like DL's but they can be challenging to style. This article is using a lot of fixed pixel widths which would break on really small screens or larger data.
Well, it took about a decade for web standards to become a real thing and a lot longer for Web Platform Tests to come to be. Still, while there are lots of tests for DOM construction and visual rendering, testing construction of the accessibility tree is lacking (also keyboard interaction testing).
And that's just for browsers, there's no shared spec for the operating system accessibility APIs the browsers' accessibility tree has to be translated into or how screen readers (and other assistive technologies) will use the OS's APIs.
I was a bit surprised to see nested <div>s given as some sort of precursor pattern, when <dl> was part of HTML before 2.0 back in the days of table layout.
I've been following roadmap.sh, and while it's not a comprehensive learning resource, it does help close obvious knowledge gaps. As it happens I was just reading about this.
That very simply puts terms side-by-side data in a nice obvious way. (Even with multiple DDs per DT.) A bit like the Wikipedia screenshot in the article but that's more balanced `grid-template-columns: 1fr 1fr;`. (But that's the flexibility of CSS Grid, right? Real easy to tweak this further for your needs/interests/design.)
The problem with 'in most cases' when it comes to a design system that's used in hundreds of different ways across departments and services, is that some week break.
I don't really like the div either (I use the design system all day, and maintain a set of components), but it makes documentation much easier.
I put dl lists in a grid with no divs needed. As MDN says, div is the last resort, invariably there is something better, and nowadays that is grid styling.
New to me is multiple dd's.
For legacy layouts littered with divs and classes, display: contents helps get rid of the div wrappers, promoting whatever is wrapped.
Even with disclosure elements there are ways to avoid div wrappers using the pseudo element for everything enclosed by the details element apart from the summary element.
I need to learn more about web accessibility, but if you completely ignore it (and other sane practices) HTML looks really simple.
I think the design of HTML is just too much. There’s so many tags that don’t do much. It’s like w3c decided that any common thing people use in websites needs a tag. The end result is more and more tags…
Can anyone convince me otherwise? It screams design red-flags to me.
PS: I love the web and think it’s the best platform and future platform we have at the moment. It’s just quirky and loves not breaking old websites!
lol, you should actually read the HTML spec, there are good explanations of all the elements. The whole point of defining semantics is that elements have meaning independent of their default appearance (or any appearance).
> I need to learn more about web accessibility, but if you completely ignore it (and other sane practices) HTML looks really simple.
Everything looks really simple when you ignore vast amounts of the subject and nuance.
Your rules don't mention keyboard or focus behavior, the only mention of either is the association between <label> and its <input>. <output> does have functionality, it's an HTML-native ARIA live region (that can be associated with a <label>).
Oh that’s great. It’s an opinionated view that focused strictly on the behavior of the tags wrt layout and appearance.
I’ve noticed that discussions of semantic meaning of tags often contain the word “feel.” Nothing wrong with that, taste matters, but it does point to the non-functional goals that are being pursued when people disagree.
<ol> vs <ul> - they are both ordered, because markup is ordered. One gets decorated differently than the other by default. Is the difference semantic or typographical?
We've always used this in our ebooks for abbreviation and glossary lists. The problem I've always had is that you need to use a bit of css to make two lined-up columns. I've done it with floats. Now, some ebook readers will support grid and flex-box, which give better results, but the Kindle still does not. Kindle is sort of the IE6 of the ebook world.
It always bugs me that the naming of the element does not seem to really fit examples like „Author: Tolkien“. It‘s not that _Tolkien“_ „defines“ the „term“ _Author_ right? The elements are still used for key-value-lists and no one seems to notice or comment on this issue.
As the article points out, HTML5 softened the name from "definition list" to "description list" to move it a bit away from thinking of it as a "dictionary" definition and somewhat better reflect how it has always been used.
> <dl aria-label="Ability Scores">
This is incorrect:
1. <dl> has no corresponding (viz. implicit) role, but can be given the role group, list, none or presentation <https://w3c.github.io/html-aria/#el-dl>.
2. You’re only allowed to define aria-label on elements that have a compatible role, implicit or explicit <https://w3c.github.io/html-aria/#docconformance-naming>.
3. aria-label is allowed on all but a handful of roles <https://www.w3.org/TR/wai-aria-1.2/#aria-label>, which in this case knocks out presentation and none, leaving group and list.
4. group doesn’t feel right, list feels acceptable.
In summary: either ditch the aria-label, or add role="list" (meaning also role="listitem" on children).
—⁂—
One thing the article misses is that you can have multiple <dt> in a row too, not just <dd>. The spec has a good example: https://html.spec.whatwg.org/multipage/grouping-content.html...
They’re not name–value pairs, they’re name–value groups.
Wow I have never noticed that, thanks for the heads up! Out of curiosity, would you put `role="listitem"` on `<div>` elements that wrap the `<dt>` && `<dd>` elements? It looks like `role="listitem"` is allowed on the `<dt>` element, but that doesn't feel like it would be accurate in the case where multiple `<dt>` elements are grouped together, and I'm not sure if that would mess with how the element is interpreted natively as as term.
I pulled up some of those spec's examples in Chrome and viewed the accessibility tree, and weirdly it assigns <dt> an implicit role of `term`, even though the spec says that <dt> should be `listitem` and <dfn> should be `term`. I didn't check FF or Safari but I would not be surprised if they did something different.
Given the spotty screen reader support for dt/dd to begin with, I'd steer clear of using multiple `<dt>` s if a11y is top priority, and just modify your content to have one dt per dd. Your single dt text could be "X or Y", repeat the definition twice, etc.
If you're familiar with Jakob's Law, it applies to a11y too, so stick with what users might have seen on other sites.
I know the fundamentals of this sort of thing, but I haven’t done much practical with it, so I don’t feel that I can comment on this point.
Your comment put me on a side quest to research the differences between i.e., e.g., viz. and sc. and I have to admit that I’m still not 100% sure
Don’t think I’ve ever heard of scilicet. <https://books.google.com/ngrams/graph?content=viz.,sc.&year_...> definitely suggests viz. to have been massively more popular than sc.
Concordantly, vis a vis, ergo..
i.e. == aka == in other words
e.g. == f.e., for example
2 replies →
The old HTML5 Doctor article is still my favourite https://html5doctor.com/element-index/
[dead]
This is going to be unpopular here, but life became easier when I quit trying to write semantic HTML. It’s just poorly designed, I’m sorry. Every time I’ve reached for a <dl> I’ve eventually regretted it because I wanted multiple levels of wrappers, or a divider between sections, or an icon, or a heading spanning multiple key-value pairs, etc. They make this stuff with some flexibility but nowhere near enough to actually cover the generalized concept it purports to. I still use the corresponding elements when there are observable benefits, of course, like <button>, <h1>, etc, but when all it’s going to do is not quite fit the data model and force me to override everything, it’s just not a practical choice.
It shouldn’t be so controversial to say that if 99% of usage routes around your API, it’s probably the API’s fault.
As a person who daily drives a screen reader, I so agree with this.
We'd all be better off if the W3 dispensed with all that ideological semantic purity BS and started doing more realpolitik. Think not about whether your API is semantically pure, but about what developers want to do, what hacks they'll use to achieve their goals despite your objections, and how to enable doing those things in a way that is maximally beneficial to everyone involved.
ARIA live regions are the perfect example. What developers actually want is `document.speakText`. What developers actually have is a weird API that announces text on the page as it changes. They have to bridge from one to the other, which is difficult and hacky, even when implemented well. But hey, at least that live region approach is semantically pure HTML...
I don't want that. I don't want to care about screen readers (unfortunately I have to). I want a system where I can pick well-defined rules and then css can style it, screen readers will understand it, automations can parse it, keyboard navigation is free.
Obviously thats not what we got, but I feel like the set of established UI patterns is manageable enough that it could be built.
A great example is the new <select> styling that developers styled in all kinds of creative ways. Now give me that for comboboxes, trees, data-grids etc...
1 reply →
Good news! `.ariaNotify()` is basically a real implementation of your hypothetical document.speakText.
https://developer.mozilla.org/en-US/docs/Web/API/Element/ari...
1 reply →
Sounds like it's CSS' fault then. I think that just like they introduced `display:contents` to remove wrappers, they should also introduce a way to group elements as if they had a common ancestor.
Proposal from 2016 (closed): https://github.com/w3c/csswg-drafts/issues/588
1 reply →
With CSS Grid math you can fake it, at least. If your DL is `display: grid;` and if you have a few extra DIVs lying around at the bottom of the DL to be borders around combined cells you just have to math which rows/columns you want to draw a border around and make the div fit that combined shape.
Good idea. Together with ::after / ::before and content: (which can insert text into the website) it might then be possible to create a website without any HTML, only CSS.
I feel similarly about HTTP. The protocol maps really well to resource stores (stuff like S3). GET, PUT, DELETE all make sense. HTTP status codes are built exactly for this use case as well. But as web developers we're mostly not developing resource stores. Those are highly generic and can be built once and used by millions of apps. Most of the time someone is writing code that interacts with HTTP they are performing RPC. You can go for GraphQL, gRPC, or many other RPC systems that just shirk the whole thing. They make everything use a POST to a single endpoint and add another layer of abstraction so you don't need to return a 4XX/5XX error for some highly application-specific situation.
It's clear the RFC writers got a little out of hand. "402 Payment Required", "407 Proxy Authentication Required", "508 Loop Detected" look to me like attempts to work in functionality specific to certain types of apps or deployments. Why do these RFC authors get their specific needs implemented into the bedrock of the web and then expect me to find where my needs happen to overlap and then tuck every aspect specific to my app into "400 Bad Request" or "500 Internal Server Error"? Every time I see a web app actually utilize more than the bare minimum of HTTP status codes I roll my eyes. Put that shit into the application layer. The protocol wasn't made for you. It was made for LAMP-stack apps serving mostly static assets.
List history (listory?) lesson, kids: As the link below to a 1985 IBM mainframe DCF/GML manual shows, DL-DT-DD have been a thing since before the web. In addition to Definition lists (DL), the 40+ year-old documentation describes Glossary lists (GL), Ordered lists (OL), Unordered lists (UL), and Simple lists (SL).
ibm :: 370 :: DCF :: SH35-0050-2 Document Composition Facility Generalized Markup Language Implementation Guide Rel 3 Mar85
https://archive.org/details/bitsavers_ibm370DCFSpositionFaci...
I believe IBM Generalized Markup Language evolved into SGML (Standard Generalized Markup Language) which was used heavily at CERN when Tim Berners-Lee was working on HTML, which is heavily derived from it. I find this interesting about HTML: some form of the markup language had been floating around for decades until Berners-Lee added hyperlinks to it.
GML dates to 1969, SGML from the 1970s. Internally we used something called BookMaster which kind of? looks like a precursor to HTML (you had :p. instead of <p>, :li. instead of <li>). There was an effort circa 1990-1991 (as TBL was developing HTML and HTTP) called HyTime which was an SGML application focused on hypermedia. HTML killed that fairly quickly.
See https://en.wikipedia.org/wiki/Charles_Goldfarb who shepherded GML/SGML and https://en.wikipedia.org/wiki/Standard_Generalized_Markup_La...
Isn't it description list?
FWIW, the Wikipedia page on "HTML elements" says:
>A description list (a.k.a. association list or definition list)[...]
The "Definition list" Wikipedia article is a redirect to that paragraph.
https://en.wikipedia.org/wiki/HTML_element#dl
The world's first website makes heavy use of <dl>s.
https://info.cern.ch/hypertext/WWW/TheProject.html
https://info.cern.ch/ (A landing page of sorts to give context and orientation about the actual first website.)
Good website. It's responsive.
> Prior to HTML5, this was called a definition list. This is because the <dl> was originally only intended to represent glossaries of terms and their definitions.
TIL I’ve been naming it wrong for a decade.
Bleh. <b> is apparently now bring attention to. As if.
Eh, it's fine, elements should be defined for what they mean, not what they look like. The explanation and distinctions made between <b> and other elements (<i>, <em>, <strong>) make sense.
The suggested (not obligatory) user agent styling for <b> is `font-weight: bolder` an agent or authors could use lots of different things to bring attention to what the element contains and treat it differently from <strong>.
https://html.spec.whatwg.org/multipage/text-level-semantics....
https://html.spec.whatwg.org/multipage/rendering.html#phrasi...
10 replies →
You’re not alone. This is the second time this week I’ve seen that, and thought it was a mistake the first time.
TIL The name was changed from a definition list.
Same here. I like definition list better ;-)
I don’t want to check what year html5 was standardized because I think it may be north of a decade ;)
> I think it may be north of a decade
Nearly two!
3 replies →
Here's a useful note on how well screen readers support DL: https://adrianroselli.com/2025/01/updated-brief-note-on-desc...
Great post. Very minor nitpick: ‘The small element must not be used for subheadings; for that purpose, use the hgroup element.’
‘The small element represents side comments such as small print. Small print typically features disclaimers, caveats, legal restrictions, or copyrights. Small print is also sometimes used for attribution, or for satisfying licensing requirements.’
(https://html.spec.whatwg.org/multipage/text-level-semantics....)
The final example of the DnD statt sheet makes me think whether it's legal to nest <dl>s?
I.e. can we do
Yes. Any flow content is permitted inside a <dd>. See:
https://html.spec.whatwg.org/#the-dd-element
https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
Plus, when curious about formal syntactic correctness, there is the validator.w3.org [1].
[1] https://validator.w3.org/nu/?showsource=yes&showoutline=yes&...
I do it to render nested JSON objects, for example.
I love DL. I think tables, at least in the past, were misused as DLs even more in the past and the inconvenience of the table markup is even worse than a bunch of divs.
It's not that inconvenient if you omit unnecessary closing tags:
I find it simpler and cleaner than any of the markdown table markups
> if you omit unnecessary closing tags
As someone who had written lots of XHTML in the past, not having closing tags makes my eyes twitch like Scrat in Ice Age. I even occasionally write `<br>` like `<br/>` out of habit.
5 replies →
Fair point, though /DT and /DD are also optional just like /TH, /TD and /TR are. So in effect, def…scription list could structurally save you one TR for each entry and two "BLE"s:
Isn't markdown table just a bunch of | ?
7 replies →
You're right, but forcing tables to cosplay as DLs was far from the worst way that tables were abused.
At least <td>s could easily centre things vertically ;)
I always thought the DL as a single row of a table.
I'm curious if the spec actually says you can only wrap it with a div because I like to do semantic html and name my elements specific to my domain.
As others have noted only the div is allowed. This isn't a unique situation either, the HTML spec despite being lenient in syntax is quite restrictive in behavior. It's unfortunate that XHTML (and XML parsing) didn't become the default as it's the opposite, more restrictive syntax, but lenient behavior.
For example in XHTML you can use custom elements as table rows or cells (provided you give them the correct role and CSS display property). This is because XHTML does not modify the tree during parsing, unlike HTML which will hoist out custom element children of the table to the table's parent.
Seems like div is the only recommended wrapper element:
https://html.spec.whatwg.org/multipage/grouping-content.html...
https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
EDIT everyone replied at once lol. I'm surprised too about div.
Also, screen reader support: https://a11ysupport.io/tech/html/dl_element
Yep, I was a little surprised about that too, seem like it is valid though https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
Yes: https://html.spec.whatwg.org/multipage/grouping-content.html...
Custom-named elements are divs.
They're not as you can see here[1] and here[2]. They both inherit from the HTMLElement interface but div is considered an HTMLDivElement which makes it distinct from a custom element.
<my-element> != <div>
[1] https://developer.mozilla.org/en-US/docs/Web/API/Web_compone...
[2] https://developer.mozilla.org/en-US/docs/Web/API/HTMLDivElem...
2 replies →
Exactly! It cracks me up when people name-check "semantic elements" when it doesn't actually mean anything in that context. Accessibility software doesn't understand the semantics of your custom elements, so there is no benefit in that situation whatsoever. Maybe it's easier for you to read and edit in the future, but that's it.
Somehow, people got convinced that <div> elements are evil and should never be used no matter what. Yes, you should use a more semantic element when it makes sense, but try to remember what that phrase actually means.
2 replies →
Aside from the accessibility benefit mentioned in the article, I would think that using semantic HTML like this would also make AI tools more effective at understanding code
I've used this a good amount of times, when I coded in front end projects. The first time gave me that satisfying feeling of using the right tool for the job, like completing a puzzle of HTML semantics. I remember JAWS not announcing it correctly in 2018, not sure if it's better now.
When I checked in 2024 or 2025, Windows Narrator announced it differently in Chrome, Firefox, Edge (Chromium mode) and Edge (IE mode), and none of them worked how I would expect them to. Adrian Roselli's verdict (https://adrianroselli.com/2025/01/updated-brief-note-on-desc...):
> Description list support continues to be generally good (with VoiceOver still the outlier), even if you may not like how it is supported.
You shouldn't try to fix this kind of thing by mangling the HTML, since (1) users tend to be used to their screen reader's quirks, and (2) in situations like these, making it juuuust right in one screen reader is likely to make it incomprehensible in another. But it is important to be aware of these quirks, so you don't accidentally design an interface that relies on less-quirky behaviour.
> Admittedly, however, support for the <dl> element is not yet universal.
Wait what? <DL> has been in HTML since.. the first draft in 1993!
I like DL's but they can be challenging to style. This article is using a lot of fixed pixel widths which would break on really small screens or larger data.
Granted, I do not know what I am doing with CSS, but the Character Sheet example seems standard flexible elements?
Some of the extracted CSS chunks
Well, it took about a decade for web standards to become a real thing and a lot longer for Web Platform Tests to come to be. Still, while there are lots of tests for DOM construction and visual rendering, testing construction of the accessibility tree is lacking (also keyboard interaction testing).
And that's just for browsers, there's no shared spec for the operating system accessibility APIs the browsers' accessibility tree has to be translated into or how screen readers (and other assistive technologies) will use the OS's APIs.
I've found CSS Grid is extremely useful for styling DLs.
I was a bit surprised to see nested <div>s given as some sort of precursor pattern, when <dl> was part of HTML before 2.0 back in the days of table layout.
It’s probably aimed at React developers, many of which are probably not even aware that elements other than <div> exist.
I loved the puzzle of thinking about the best semantic representation of content in HTML. Good times
I've been following roadmap.sh, and while it's not a comprehensive learning resource, it does help close obvious knowledge gaps. As it happens I was just reading about this.
https://roadmap.sh/html (see "Definition lists")
Hoped to see CSS for the alternative, where <div> is not nested inside the <dl>. Too used to thinking of div as "layout containers."
CSS Grid is a very good friend when trying to style DL without extra DIV wrappers. An example:
That very simply puts terms side-by-side data in a nice obvious way. (Even with multiple DDs per DT.) A bit like the Wikipedia screenshot in the article but that's more balanced `grid-template-columns: 1fr 1fr;`. (But that's the flexibility of CSS Grid, right? Real easy to tweak this further for your needs/interests/design.)
The GOV.UK Design System summary list component is a description list https://design-system.service.gov.uk/components/summary-list...
And... it also uses the wrapper div for styling
The wrapper div is making me a bit sad. These days, using grid layout, you don’t actually need it in most cases
The problem with 'in most cases' when it comes to a design system that's used in hundreds of different ways across departments and services, is that some week break.
I don't really like the div either (I use the design system all day, and maintain a set of components), but it makes documentation much easier.
Absolutely!
I put dl lists in a grid with no divs needed. As MDN says, div is the last resort, invariably there is something better, and nowadays that is grid styling.
New to me is multiple dd's.
For legacy layouts littered with divs and classes, display: contents helps get rid of the div wrappers, promoting whatever is wrapped.
Even with disclosure elements there are ways to avoid div wrappers using the pseudo element for everything enclosed by the details element apart from the summary element.
The <dl> tag seems to cover a subset of a broad semantic space, but doesn’t easily extend beyond adding another <dd>.
I dunno, I guess I’m a caveman. If it looks right and works (including accessibility) then I figure I’m pursuing something that doesn’t matter a lot.
Shameless plug: you might be interested in https://calvinlc.com/p/2026/02/11/everything-is-a-div.html
I need to learn more about web accessibility, but if you completely ignore it (and other sane practices) HTML looks really simple.
I think the design of HTML is just too much. There’s so many tags that don’t do much. It’s like w3c decided that any common thing people use in websites needs a tag. The end result is more and more tags…
Can anyone convince me otherwise? It screams design red-flags to me.
PS: I love the web and think it’s the best platform and future platform we have at the moment. It’s just quirky and loves not breaking old websites!
lol, you should actually read the HTML spec, there are good explanations of all the elements. The whole point of defining semantics is that elements have meaning independent of their default appearance (or any appearance).
> I need to learn more about web accessibility, but if you completely ignore it (and other sane practices) HTML looks really simple.
Everything looks really simple when you ignore vast amounts of the subject and nuance.
Your rules don't mention keyboard or focus behavior, the only mention of either is the association between <label> and its <input>. <output> does have functionality, it's an HTML-native ARIA live region (that can be associated with a <label>).
https://html.spec.whatwg.org/multipage/
Oh that’s great. It’s an opinionated view that focused strictly on the behavior of the tags wrt layout and appearance.
I’ve noticed that discussions of semantic meaning of tags often contain the word “feel.” Nothing wrong with that, taste matters, but it does point to the non-functional goals that are being pursued when people disagree.
<ol> vs <ul> - they are both ordered, because markup is ordered. One gets decorated differently than the other by default. Is the difference semantic or typographical?
5 replies →
All the native GUI toolkits are dead, but people can write entire essays now on a single HTML element. I guess that's progress?
What about multiple '<dt>' for one or more '<dd>'?
We've always used this in our ebooks for abbreviation and glossary lists. The problem I've always had is that you need to use a bit of css to make two lined-up columns. I've done it with floats. Now, some ebook readers will support grid and flex-box, which give better results, but the Kindle still does not. Kindle is sort of the IE6 of the ebook world.
I loved the character sheet example! Fun!
This seems a clear enough win for things that would fit into a simple python dictionary.
Why is it preferred over <table> for laying out columns via a the character attributes at the bottom of TFA?
blog looks beautiful. I really wish I had this kind of talent for frontend.
it's on archive html5 .flac 16-bit 44.1kHz no <dl> flag.
It always bugs me that the naming of the element does not seem to really fit examples like „Author: Tolkien“. It‘s not that _Tolkien“_ „defines“ the „term“ _Author_ right? The elements are still used for key-value-lists and no one seems to notice or comment on this issue.
Am I the only one?
As the article points out, HTML5 softened the name from "definition list" to "description list" to move it a bit away from thinking of it as a "dictionary" definition and somewhat better reflect how it has always been used.
[dead]
Good title
[dead]
[flagged]