A missed opportunity to not have all of these examples inline. The page/blog-post would be so much more convincing if it utilized all of these HTML replacements instead (or in addition) to linking to codepen.
Absolutely mind boggling. I've seen it many times before. There's some FooMaker v1.0 announced and you click on it thinking it will allow to easily make Foo and the example is how to disable FooMaker's lights l while AC is running or some other obscure edge case that affects like one in a million people, no examples how to actually use it for most common use case or how the result look like.
The details / summary thing absolutely kills me. There’s basically nothing you can’t do with them. Hiding and replacing markers is easy. But every component library just pretends they don’t exist.
It even saves you the effort of all the aria control and expanded tags: these tags don’t need them.
One drawback of details was that cmd+f search wouldn't play nicely when the details was closed. But now there's a hidden="until-found" you can put on child content, along with an associated event. So you can open the details when a user searches :) super useful
You don't need the hidden="until-found" for details/summary, because that has those semantics automatically, but you can use that for other elements that behave similarly (for example tabs, which can't quite correctly be implemented with details/summary, and so needs to be done by hand).
Also I think the event isn't currently emitted consistently on all browsers (and maybe not at all for hidden="until-found"?) so unfortunately you can't quite rely on that yet if you need to sink some JavaScript state to your html. But in general, yeah, this is a really cool feature.
That is no longer true! You can do it in CSS with a combination of `@starting-style` and `transition-behavior: allow-discrete`. [1]
Another gotcha you'll run into is animating the height. A couple other new features (`interpolate-size: allow-keywords` and `::details-content`) will let you get around that. [2]
The major issue with this is that modern CSS is almost its own job, to the point we used to have Interface Developers at some place I’ve worked (HTML+CSS specialists). I did frontend for over a decade and eventually lost the train on CSS changes, I don’t even know what’s going on there anymore.
It’s still awesome, but it’s becoming increasingly silly to ask someone to know modern HTML, CSS, JavaScript, Typescript, some build tools, a couple of frameworks, etc.
The amount of JS we ship to clients is a reflection of cost-cutting measures at your workplace, not that every FE dev shuns CSS.
Is that "straight-forward easy to understand and follow JavaScript" the whole thing written from scratch? Or does it use libraries (that use libraries, that use libraries)?
Because I've written my share of javascript-from-scratch in my time - before npm and such. And even if my use-case was limited, in order to get edge-cases and details working - issues long solved by their HTML/CSS counterparts - we needed more and more JS. Many of which handwritten polyfills, agent-detection, etc.
Seriously, things like scrollbars (because the client insisted on them being consistent across user-agents) or dropdowns (because they had to be styled) "visited" state on links, in pure JS are thousands of lines of code.
Maybe not today, anymore, IDK, with current APIs like the history API or aria labeling. But back then, just in order to make the dropdown work with screen readers, or the scrollbars react well to touchpads -in the direction the user was used to based on their OS- took us thousands of lines of JS, hacks, workarounds and very hard to follow code - because of the way the "solutions" were spread out over the exact right combination of JS, HTML and CSS. Edit: I now recall we got the web-app back with the comment "When I select "Language" and start typing "Fr" I expect French to be picked and "enter" to then put the language in French". We spent another few days on functions that collect character inputs in memory and then match them with values. All because "flags in front of the names were of crucial importance".
So, maybe this is solved in modern HTML/CSS/JS. But I highly doubt it. I think "some straight-forward ... JavaScript" is either an `import { foo } from foobar` or a pipe-dream in the area of "programmers always underestimate hours"
Because you need 20x the JS to do the same thing and it’s still not hardware accelerated. These new CSS properties are well supported and will only get better.
Because a team of browser engineers have already written and reviewed the code to do it for you; and (hopefully) it’ll be performant, properly tested and accessible… ;-D
JS animations run on the main thread with everything else, so if your browser is busy doing something else the animation ends up being janky. Using CSS solves that problem.
I agree, but must also observe that I have never met a designer who was willing to admit without a knock-down drag-out fight that any animation they put in was not somehow crucial.
I've never met a designer who wasn't completely fine with my suggestions for more pragmatic solutions. Like just styling a default scrollbar instead of implementing my own scrollbar to make it exactly like the design. Using a default drop-down menu instead of rolling my own just so I can round the corners of the selects.
The designers I've worked with are fine with these things. We have more important things to work on than small style details. We can go back and change these things later if anyone actually cares, but generally nobody ever does.
You're doing it wrong. You don't have to remember the incantations. You just have to remember that they exist, and then google them or ask an LLM when you need them.
If you use something enough you'll remember. If you don't, you just look it up when you need it. This is basic programming, nobody remembers everything.
Seriously. As a user I can count on zero hands the number of times I’ve said “Oh great, I’m sure glad this UI is animated!” - and likewise zero times have I missed it when animation isn’t used. Animation is a way to light small units of your users’ precious time on fire, for zero benefit.
As the other user alluded to, Animations are not actually there for people who are comfortable using a computer. The vast majority of users are borderlines in capable of using the internet these days. Animations are supposed to be there to really help guide these users into understanding what the scary machine is doing when they click it. Can they be overused, absolutely, but i think have an accordion fold out animated is a reasonable case. You gotta remember your average user isnt paying any fucking attention, so drawing their attention to important changes on screen is not only good but necessary. I'd prefer no animations ever, but i also dont own an iphone while the majority of the world either does or wants to.
That's the positive interpretation, but none of the discussions I've had with UI designers or managers have been about adding animation for accessibility, and the zeitgeist of the last decade has been that skeuomorphism (of which intuitive animations are a subset) is passé.
So far as I can tell, all that the stakeholders want from the UI, animations included, is pizzazz.
Animations are also a way to explain causal relationships between interactions and their results, and to help build mental models of software behaviour.
Being related to neither software behavior nor the structure of the underlying problem, animations tend to obscure the causal relationships and make it harder for user to build a correct mental model.
> Animating the details element is tricky. By the spec, browsers don’t natively support transitions between display: none and display: block.
Very hot take; then don't animate them!
Animation in a UI is great - you draw the user's attention to a widget that changed because they might not necessarily notice it otherwise. This improves the UX.
With a details/summary, the animation is not needed and can only make a negative change to the UX. There is no positive change to the UX that animating the details/summary elements would bring. When it is opened it is obvious.
If you really really need to animate the details, instead of animating open/close, instead animate the summary background/text color to indicate that the element has just changed state.
Would I like easy animation of open/close? Sure. Does it improve the UX? Nope.
Fun fact: <details> even works on github and similar sites with markdown-based input. You can post large inline logs in issues without cluttering the conversation.
Details works even when it's set display:contents too, for even more flexibility. Can't animate from open›close, yet, though. That's pretty much my last frustration with it.
I think the CSS support for that has finally landed, though it means targetting a pseudo element instead. Its been a year, so support is probably good enough you don't care if just the animation doesn't happen.
Note that the transition to `auto` in that post relies on `interpolate-size` which has yet to land on Firefox or Safari, and neither have movement right now.
Yes. For example, on Codidact (https://codidact.com), limited HTML access is offered along with Markdown when making posts, and the details and summary tags in particular are whitelisted. I've made extensive use of this in some of my content, for example https://software.codidact.com/posts/289251/289252#answer-289... . If you have NoScript you can easily verify that the expanding sections work perfectly well without JavaScript. They can even be nested, as they are here; and the summary text can contain some other forms of markup without issue. (When crafting the post, however, I had to do some tricky things with whitespace to avoid confusing the Markdown parser.)
You can't actually control the open state properly from markup (the `open` attribute only sets the default state), which is why I haven't bothered with them.
I’m not sure this is correct. The DOM class HTMLDetailsElement has the `open` property, which you can use to read/write the details element’s state. If you’re using setAttribute/getAttribute just switch to the property.
Having to use the property on the element instance, rather than the actual HTML attribute, is exactly the kind of wrapper code I want to avoid if I'm using a built-in.
You need some JS to change an attribute as much as you need JS to change a property. What am I missing?
I hope the command attribute (https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...) will eventually support this out of the box. In the meanwhile you can write a single custom --toggle command which is generic and works with any toggleable element
`open` works just like checked for a checkbox input. You can set the initial state in HTML and CSS can respond to changes in the attribute if a user clicks it. Markup won't have programmatic control over the attribute by design, that's always done in JS by modifying the element instance.
If you specify the same name on each `details` element they behave like an accordion automatically [1], no need for JavaScript. If you set one of them to open that one will be initially open.
The interesting part here isn’t “no JavaScript”, it’s that HTML already covers more use cases than people remember (forms, dialogs, validation, navigation).
I ran into this repeatedly while writing my book "You Don’t Need JavaScript"[0]: most JS in these cases isn’t adding capability, it’s compensating for forgotten platform features.
Agreed! I assume the reason for the forgetting of the features is that at least some were poorly supported when first released so developers create workarounds that then become the de facto standard.
It has been amazing to see the speed up in release and support of new CSS features over the last couple of years! Even the masonry layout has finally reached an experimental stage
Yup, at this point it feels more like habit than necessity. People learned to build things like dropdowns in JavaScript years ago, so they keep doing it that way.
A lot of devs simply don’t look any further when it comes to what HTML and CSS already provide.
Most of this is great, except for the input/datalist bits, which are not sufficiently functional to be used in any real scenario. Users expect these interfaces to be tolerant of misspellings, optional sub text under each option, mobile ux niceties, etc -- and so everyone builds this with js...
My main beef with datalist is that there's no easy way to show and allow only text (e.g. Beverly Hills), but have the actual value selected be a number (e.g. 90210). In other words there's no analogy to <option value="90210">Beverly Hills</option>.
That and there's no HTML way to interactively load results. Or are you really going to serialize half a million records to HTML and transfer it all every time the relevant block is added to a page? What if it sits in the header or footer templates?
Right, I sort of expected there to be an attribute for an url.
<datalist json="search.php?q=toyota+corolla">
But then you would want to send other form values along with it which might make things more complicated than it should be?
Static could be better too. When search engines first started building these auto complete dropdowns the multi word input was really the killer feature. To have something like "green toyota" you would have to generate an element for all color and brand combinations? And the you want it to work for "green toyota corolla" and you get an abc kind of list length.
Perhaps a wildcard would have been fun or regex options.
> Each <option> element should have a value attribute... It can also have a label attribute, or, missing that, some text content, which may be displayed by the browser instead of value (Firefox), or in addition to value (Chrome and Safari)... The exact content of the drop-down menu depends on the browser, but when clicked, content entered into control field will always come from the value attribute
This seems... underspecified. Not ideal that Chrome/Safari aren't aligned with Firefox here, and that there is no standard way to only display the label
Just been through several frontend interviews in the last few months, where it's clear that they still judge a developer's JS skills (especially React) than being semantically correct on HTML elements.
Every question/exercise is centred around how well you know React hooks, effect, memoization, modern css-in-js etc. Given I've been working with Astro recently, in one interview I talked about DOM APIs and I can see the interviewer raise an eyebrow. In later stage, even I that passed the exercises, still didn't get the job.
Remember that a large part of hiring is finding someone who fits in an existing team. A team that uses react won't appreciate someone choosing to use native DOM APIs instead of a react component.
Eh. I build apps with Preact, but I prefer candidates who know the core web platform. They’ll be more apt to use the right tool for the job and not be baffled by edge cases.
One thing I am quite hopeful for is customizable selects! It's in WHATWG stage 3 right now. I have seen so many horrors with javascript-based custom dropdowns components. https://developer.chrome.com/blog/a-customizable-select
Plain HTML is very cozy to me - I came of age in that era. Marquee tags 4eva.
But as much as I hate to admit it, it is very difficult to build something functional today with plain HTML and no/minimal JS. If you want, say, a model form that manages its children as well, you're basically going to end up with a 2003-era ASP-feeling application with way too many views and forms (as seen on your employer's current HR system). Or you use HTMX... and you still end up with just as many (partial) views, but now with so much implicit state that you're veering into write-only code.
I dislike modern JS to the extent that I opted for Phoenix LiveView, just so I could achieve interactivity without ever having to touch JS, but in truth it's not a comprehensive solution. Still had to write a web worker, a bridge to handle things like notifications, etc. Plus the future direction of Phoenix, all in on AI, is worrying.
Honestly, I should probably just swallow my disdain and learn to appreciate and use modern JS, as painful as that sounds. I want to write and release cool things, not get caught up in navel-gazing language wars.
> But as much as I hate to admit it, it is very difficult to build something functional today with plain HTML and no/minimal JS.
I would certainly agree that using a little JS can get you further than just HTML. But I think that a plain HTML page is far more pleasant to use (and thus, functional) than the JS monstrosities that dominate the Web today. There's a reason people use the NoScript addon: because a whole lot of website designers use JS in ways that make the experience a ton worse for the user.
It's not an either/or. Modern Javascript is actually really nice to write and use, and you can write it in a tight, minimal way that doesn't bloat the page or slow it down.
Of course you can, but most people still opt to pull in a whole framework (React) or heavy library (jQuery) just to achieve what's essentially a few XMLHttpRequests and some DOM changes.
> Or you use HTMX... and you still end up with just as many (partial) views, but now with so much implicit state that you're veering into write-only code.
You're overthinking htmx then. I do some fairly complex stuff with no extra partials. Trick is just always rerender and use hx-select and hx-target to slice out the bits you want to update on the current page.
Server always has authoritative state and code is dead simple to reason about.
> I do some fairly complex stuff with no extra partials. Trick is just always rerender and use hx-select and hx-target to slice out the bits you want to update on the current page.
Good trick! My only experience of HTMX in production entailed porting Stimulus code, hence the partials, but your approach is obviously much neater. I'll give it a shot, next time it might be suitable.
OT: marquee tags were a missed opportunity to implement horizontal scrolling often used on shopping websites. Now it uses JS to achieve the same.
I have been trying to find other more commonly known UI patterns that could be done natively. The time has long come for tabular data to be put into HTML tables just by referencing the source. Xslt almost did that. Another one is integrating xml http requests with native html. I think HTMz came close to this.
There's a usability and design issue with that as you lose what you're reading as it scrolls off the screen. Also, scrolling is a styling issue and not a document description issue which is what HTML is for.
Note: <marquee> has never been part of any HTML standard since the beginning except the current one which only has it for the purpose of marking it obsolete so people will quit using it.
I appreciate the helpful reply, but I think this is precisely the kind of indirection I need to avoid. I'm a sucker for elegance. If left entirely to my own devices I'd probably design a language / write a transpiler of my own, and wind up with exceedingly elegant tooling for websites, and no websites. :)
This is why I don't use Typescript or frameworks in my own projects, I just constantly seek the cleanest abstraction and never get anything done. Using a deliberately messy solution is annoying but at least I accomplish stuff.
Before anyone wastes a lot of time like I did trying to use the popover API: it is not ready yet. You can do very basic things in all browsers but positioning is still different and/or totally broken per browser.
HTML and JavaScript serve distinct purposes, making better or worse comparisons logically flawed. Complex/interactive web apps requires JavaScript, period. Attempting to build sophisticated apps solely through HTML (looking at you HTMLX) eventually hits a functional ceiling.
I dont think anyone is arguing Google Earth should be pure HTML. But it is equally false you cant do Gmail with HTML only.
There are things that HTML could do, and should be doing, that is not done or not yet possible simply due to hype and trend from browser vendors. We could continue to polish HTML + sprinkle of Javascript to its absolute maximum before hitting JS Apps. Right now this is far from the case.
Hey the email services has just proved you could offer better than Gmail experience with HTML + small dose of JS. Another example being the new FE on Github.
At the end of the day it isn't really the tech that is the problem. Is how people use the tech. And for thousands of different reasons keeping it simple has always provided better experience evaluated on the whole.
Github's old frontend was mostly HTML with a bit of JS, their new frontend is react. The old UI had its bugs, but it was much better than the react version in my experience. I still commonly find the UI out of sync with itself requiring a reload, but now I also frequently wait for the page to load and viewing large diff's is a performance nightmare.
I assume you mean htmx. It doesn't have to be either/or. You can supplement htmx with Javascript.
The core idea with htmx is that you transfer hypertext with controls and structure built in, not just a JSON blob that requires additional context to be useful.
I have just shipped a very useful and interactive app surprisingly quickly for my customer using just htmx with a little Javascript.
It shouldn't have to be this way though. There is no reason html can't do things it needs to do to build complete apps. We could use reasonable defaults to allow a new type of html markup without JavaScript.
All the http verbs.
Decent html input controls
What else?
Depends on how complex it is meant to be. Just like many wordpress sites that could easily have been static websites, many javascript heavy sites could have easily just have been using htmlx.
If your need really, goes beyond what htmx offers, then you may need Javascript. But in my experience people tend to use the tools they know for their job, not the tools that would be best suited.
FYI, it's easy to cache the html output of a WordPress site, resulting in essentially a static site with graphical admin, page builder, and all the other bells and whistles.
I don't want to be so negative, but we got details and popover after literally decades, and we still have datalist presented as a plausible option? Html is so underdeveloped and the first we all agree on that the first we'll pretend from committee real advancement
i like the points the article makes, but i really wish it used looping videos instead of actual GIFs
i don't really see any reason to use GIFs here; any widely available video codec like H.264, VP8/VP9 or AV1 will result in significantly smaller file sizes, look better, and will allow enabling controls for seeking and play/pause
It's great until you have a typo in the field, or want to show options that don't start with what you typed in but appear near the end of an option (think Google search's autocomplete). There's no way to filter in Javascript and force it to show certain options with <datalist>. I've resorted to <ol> for search suggestions.
I don't see any mention of HTTP 204 or multipart/x-mixed-replace. Those are both very helpful for implementing rich JavaScript-free HTML applications with advanced interactivity.
Until your client tells you that it doesn't work in Edge and you find out it's because every browser has its own styling and they are impossible to change enough to get the really long options to show up correctly.
Then you're stuck with a bugfix's allotment of time to implement an accessible, correctly themed combo box that you should have reached for in the first place, just like what you had to do last week with the native date pickers.
I didn't know about <datalist>, but how are you supposed to use it with a non-trivial amount of items in the list? I don't see how this can be a replacement for javascript/XHR based autocomplete.
> If we can hand-off any JS functionality to native HTML or CSS, then users can download less stuff, and the remaining JS can pay attention to more important tasks that HTML and CSS can't handle (yet).
Brilliant I will be adopting a few of these, I have been on a personal quest to reduce js use recently, I feel like I spend more time debugging js than producing the end result.
I was trying to rewrite some UI library with html sometime ago following the W3C accessibility specs and found out a lot of patterns can’t be done with pure html and require javascript unfortunately.
If we've concluded that's it's okay to have elements that change/morph, as we seem to with the introduction of things like details, a native tab-like element feels like a glaring omission. Tabs have been a long-standing UI pattern and forcing every site to implement their own is a nightmare for accessibility. (The page you're reading is maybe already in a browser tab.)
I wouldn't be surprised if it turned out less than half of the custom tab interfaces on the web failed from an accessibility standpoint. When considering ARIA guidance, I don't even think it's possible to build an accessible version in HTML alone.
Other people have recognized it's missing. Open UI has a draft spec for it[0] and CSS Tricks has an article from 2001 about Open UI's experiments with sections for tabs[1]. I have no idea what happened on this front, though.
Accordion behavior is discussed in the article in the "Accordions / Expanding Content Panels" section:
> Use the same name attribute on all related details (like radio buttons) to restrict only one open panel at a time
And tabs can be a <details>-based accordion with some creative CSS to adjust the layout (left as an exercise for the reader, but I could write up an example if that would be helpful!)
And yes, being able to do all of these in pure HTML/CSS would be awesome. Though we are getting there with things like `details` and the newer `popover` features which should make things like rich tooltips, menu buttons, etc. a lot easier to implement. IIRC, there are also several anchor CSS properties to make positioning a lot simpler.
Use a checkbox, d. Define vars for light mode. Override when checked for dark mode with body:has(#d:checked) and can include the dark mode media query too
Why would you build a switch instead of relying on the user’s system settings? The only reason I can imagine is that your dark/light mode is not usable/readable so it forces the user to switch
I often use different light/dark settings between apps and my system. Just because I want system UIs to be dark, for example, doesn't mean I want to read long pages of white-on-black prose on your blog.
Seconded. Just because I like to have the browser toolbar dark and GitHub dark doesn't mean I also want to read lengthy articles (LWN) in thin white text on a black background.
Having it default to the users preference is nice, but you should still provide an override. I sometimes use my browser in light mode while my OS is dark mode. Many times, I find the contrast for dark mode websites too low unless I’m in a totally dark room.
Your blogs have very small amount solution, but the JS use cases are very large. How this little replacement can do more thing? I usually like the idea of being using as lean as possible, if it's can be possible to do more thing just with HTML and CSS that's obviously cool. Is it really possible to replace JS with HTML in near future?
BTW the toggle solution (expanding content) is good.
A pleasant surprise to see Aaron's post here, we worked together for a bit on frontend optimization in a multi-tenant international ecommerce platform. That work was a large part of my inspiration for building https://contentblocksjs.com which encapsulated a lot of the JS concerns into web components.
That’s a bit of evergreen topic. “stop bloating web with js” comes up fairly often and there are those people who think they found a solution and everyone should start using whatever they imagine is “best for everyone”.
In my opinion most of those people struggle with whatever they encountered in ecosystem and just want to find a way that fits them - while also trying to make others do the same.
*“You didn’t want to make things perfect. You just hated things the way they are.”*
Something I keep thinking about when I consider the trade-offs between building a site with HTML/CSS wherever possible vs JS is what the actual _experience_ of writing and maintaining HTML/CSS is vs JS. JS gets knocked around a bunch compared to "real" languages (although less so in recent years), but at the end of the day, it's a programming language. You can write a loop in it.
Writing a web server in C++ is a way to get excellent performance. So why don't most people do it?
JS got popular because some devs were trying to realize a world where the same code can be shared on the front and backend. I think on the surface, it was a noble goal with good intentions. Having only one programming language to handle is going to create some efficiency gains when you work in a large company with thousands of devs who all need training want to share knowledge as a larger organization.
In the past decade, we went full JS as an industry and now we’re starting to swing back. Server side interactivity like Phoenix Liveview, C# Blazor, HTMX, PHP/Laravel Livewire, Rails Hotwire, all of these are different abstractions around JS to make interoperability between the frontend and backend more manageable and they’ve come a long way to closing the gap. Advancements in HTML/CSS standards also deserve credit for closing the gap but we’re still not quite there yet.
But at the end of the day, the web is dynamic. As new tools and techniques are discovered, the industry will continue to evolve and certain “hacks” will become new standards and ignorant newcomers will reinvent the wheel again to achieve some crazy interactive design because they didn’t know any better! And it wil work, mostly.
Until the way we interact with browsers changes, I feel that we’ll continue to bolt on new features over time and the web will continue to evolve. Just like the iPhone, a surge of use of smart glasses could change the computing paradigm or perhaps its some other device entirely.
So you can (and should!) try to optimize for today, but trying to optimize for tomorrow will always carry the risk irrelevance if the market pivots quickly. Bleeding edge is risky but so is falling behind.
Some of these new HTML features don't fully work in my "ancient" browser. But all of them partially work (ie opening the accordion element doesn't close others but it still opens and closes) and they still remain functional elements I can read and interact with. This puts them far ahead of any javascript implementation which almost universally fail to nothing.
I really try not using JavaScript unless absolutely needed. On my latest project, the whole site actually functions without JavaScript and is server side rendered. However, there's some small piece which I really needed JavaScript for couple reasons.
Basically, I have a site which collects the top STEAMD posts from places like HN, lobsters, tildes, slashdot, bear, reddit etc and displays them in chronological order. I wanted a way for users to block posts with certain keywords or from specific domains. I didn't want to do this server side for both performance reasons plus privacy reasons. I didn't want users to need signing up or something to block. I also didn't want to collect block lists for privacy reasons. So, I resorted to using JavaScript and local storage. All posts within the filter for the date are sent and JavaScript is used to block posts with keywords before displaying. So my server never knows what keywords are blocked.
I'm so not impressed by the toggle implementation... How nice it could have been.
Nesting the elements is a truly hideous choice. The summary is part of the details?? I thought they were opposites.
Should we also put the headings in the <p> from now on?
Identifying a target should be done by id or by name. That it does use a name because js can't target it without makes it even more stupid.
We already had labels for form fields. Inventing a completely different method for something very similar is a dumb idea. The old checkbox hack is more flexible and less ugly for some implementations.
Why force the hidden content to be below or above the toggle? We aren't gaining anything with this.
What is this nonsense for an element to not just be hidden or displayed but to have some weird 3rd state where only one of its children is shown?
How should styling it even work for this new state? If I apply a style to the hidden content it must also apply to the link? The text is hidden but the style is visible??? Preposterous!
Don't try style <details> to avoid unexpected behavior. Try wrapping the hidden content in a new element to make it behave normally.
What is this ugly arrow? If you find 1000 websites using a toggle I doubt there is one using an ugly arrow like that.
The default styling gives no clue about it being clickable?
The pointer (awkwardly called the cursor) choice is the text selection?????
Blue underlined "more" is what everyone does and everyone is used to. The cursor should be pointer. (This is css speak for "the pointer should be a hand")
The number of js toggles you can find online where the button lives inside the hidden text is guaranteed to be zero. Forget about drop in replacement, you will have to reinvent your css.
Maybe I'm dense but I also want my url to reflect the state of the page. I would have been impressed if that was supported. Personally I use actual links and disable default action in the listener if js is enabled/working or modify the state on the server if js isn't available/working.
It would have been great if the toggle action was implemented as a simple attribute something like toggle="element name" so that anything can be clickable and anything can be toggleable. Have a "closed" as well as an "open" attribute for the target.
Doesn't seem very hard. An open/closed attribute would be useful for other things too. Using display:none is terrible as display: is used for many things.
> Nesting the elements is a truly hideous choice. The summary is part of the details?? I thought they were opposites.
It gives them a semantic connection. Last I checked, HTML isn't really based on giving special meaning to combinations of sibling tags. A summary is part of the thing that conceptually requires detailing.
> If you find 1000 websites using a toggle I doubt there is one using an ugly arrow like that.
I think the default looks fine. But TFA clearly explains right there that it can be styled. (Specifically, by styling ::before on the summary tag.)
> The default styling gives no clue about it being clickable?
You asked what the arrow is, and then asked about the lack of indication that the summary header is clickable. The arrow is exactly that indication.
> Maybe I'm dense but I also want my url to reflect the state of the page.
If you scroll, should the fragment automatically update as you scroll past anchors? I think I'd find that quite annoying.
The problem is that it's difficukt to style or animate those things. Unless you're builsing something for dun or technical where it's not important it's fine but i doubt any real world commercial project would be satisfied with just this
A missed opportunity to not have all of these examples inline. The page/blog-post would be so much more convincing if it utilized all of these HTML replacements instead (or in addition) to linking to codepen.
Absolutely mind boggling. I've seen it many times before. There's some FooMaker v1.0 announced and you click on it thinking it will allow to easily make Foo and the example is how to disable FooMaker's lights l while AC is running or some other obscure edge case that affects like one in a million people, no examples how to actually use it for most common use case or how the result look like.
The details / summary thing absolutely kills me. There’s basically nothing you can’t do with them. Hiding and replacing markers is easy. But every component library just pretends they don’t exist.
It even saves you the effort of all the aria control and expanded tags: these tags don’t need them.
One drawback of details was that cmd+f search wouldn't play nicely when the details was closed. But now there's a hidden="until-found" you can put on child content, along with an associated event. So you can open the details when a user searches :) super useful
You don't need the hidden="until-found" for details/summary, because that has those semantics automatically, but you can use that for other elements that behave similarly (for example tabs, which can't quite correctly be implemented with details/summary, and so needs to be done by hand).
Also I think the event isn't currently emitted consistently on all browsers (and maybe not at all for hidden="until-found"?) so unfortunately you can't quite rely on that yet if you need to sink some JavaScript state to your html. But in general, yeah, this is a really cool feature.
Oh huh I didn't know details supported that naturally. I'm guessing this wasn't always the case and my knowledge is simply outdated. TIL!
When you search text with ctrl+f it also searches inside details elements and automatically expands them!
> The details / summary thing absolutely kills me. There’s basically nothing you can’t do with them.
Animating the details element is tricky. By the spec, browsers don’t natively support transitions between display: none and display: block.
That is no longer true! You can do it in CSS with a combination of `@starting-style` and `transition-behavior: allow-discrete`. [1]
Another gotcha you'll run into is animating the height. A couple other new features (`interpolate-size: allow-keywords` and `::details-content`) will let you get around that. [2]
Modern CSS is awesome.
[1] https://developer.chrome.com/blog/entry-exit-animations
[2] https://nerdy.dev/open-and-close-transitions-for-the-details...
The major issue with this is that modern CSS is almost its own job, to the point we used to have Interface Developers at some place I’ve worked (HTML+CSS specialists). I did frontend for over a decade and eventually lost the train on CSS changes, I don’t even know what’s going on there anymore.
It’s still awesome, but it’s becoming increasingly silly to ask someone to know modern HTML, CSS, JavaScript, Typescript, some build tools, a couple of frameworks, etc.
The amount of JS we ship to clients is a reflection of cost-cutting measures at your workplace, not that every FE dev shuns CSS.
I can't see how a bunch of esoteric incantations are better than just some straight-forward easy to understand and follow JavaScript.
Is that "straight-forward easy to understand and follow JavaScript" the whole thing written from scratch? Or does it use libraries (that use libraries, that use libraries)?
Because I've written my share of javascript-from-scratch in my time - before npm and such. And even if my use-case was limited, in order to get edge-cases and details working - issues long solved by their HTML/CSS counterparts - we needed more and more JS. Many of which handwritten polyfills, agent-detection, etc.
Seriously, things like scrollbars (because the client insisted on them being consistent across user-agents) or dropdowns (because they had to be styled) "visited" state on links, in pure JS are thousands of lines of code. Maybe not today, anymore, IDK, with current APIs like the history API or aria labeling. But back then, just in order to make the dropdown work with screen readers, or the scrollbars react well to touchpads -in the direction the user was used to based on their OS- took us thousands of lines of JS, hacks, workarounds and very hard to follow code - because of the way the "solutions" were spread out over the exact right combination of JS, HTML and CSS. Edit: I now recall we got the web-app back with the comment "When I select "Language" and start typing "Fr" I expect French to be picked and "enter" to then put the language in French". We spent another few days on functions that collect character inputs in memory and then match them with values. All because "flags in front of the names were of crucial importance".
So, maybe this is solved in modern HTML/CSS/JS. But I highly doubt it. I think "some straight-forward ... JavaScript" is either an `import { foo } from foobar` or a pipe-dream in the area of "programmers always underestimate hours"
Because you need 20x the JS to do the same thing and it’s still not hardware accelerated. These new CSS properties are well supported and will only get better.
Because a team of browser engineers have already written and reviewed the code to do it for you; and (hopefully) it’ll be performant, properly tested and accessible… ;-D
JS animations run on the main thread with everything else, so if your browser is busy doing something else the animation ends up being janky. Using CSS solves that problem.
Is CSS too hard to learn or something?
1. CSS is declarative, so it avoid subtle bugs
2. CSS integrates better with HTML, as it has selectors to automatically bind to elements (yes there custom elements for JS)
@starting-style Has less than 90% browser support making it a non-starter for the time being at least.
It'll just degrade gracefully into not animating the element's entry, so unless the animation is somehow crucial you should still be fine to use it.
If you really need to detect whether it's supported there are hacky methods: https://www.bram.us/2024/07/11/feature-detect-css-starting-s...
I agree, but must also observe that I have never met a designer who was willing to admit without a knock-down drag-out fight that any animation they put in was not somehow crucial.
The trick is they'll see it working for themselves. :)
I've never met a designer who wasn't completely fine with my suggestions for more pragmatic solutions. Like just styling a default scrollbar instead of implementing my own scrollbar to make it exactly like the design. Using a default drop-down menu instead of rolling my own just so I can round the corners of the selects.
The designers I've worked with are fine with these things. We have more important things to work on than small style details. We can go back and change these things later if anyone actually cares, but generally nobody ever does.
91% of usage for browsers tracked by caniuse [1].
The biggest gap is Chrome versions > 2 years old.
[1] https://caniuse.com/?search=%40starting-style
You say awesome, I say layers upon layers of opaque incantations I have to remember. Thank god for LLMs.
“Layers upon layers of opaque incantations”
You’ve described software.
You're doing it wrong. You don't have to remember the incantations. You just have to remember that they exist, and then google them or ask an LLM when you need them.
If you use something enough you'll remember. If you don't, you just look it up when you need it. This is basic programming, nobody remembers everything.
And how might you do this in Javascript?
they just said "LLMS"!
My point was that js would be vastly more complicated than these html/css "incantations".
Most front end engineers could do it in JS without ever having to look something up. But the CSS to do it is still obscure to most.
Animating accordions is almost always a bad idea because the content length can make it unbearable.
In general I find animations on the web overused and unnecessary
Seriously. As a user I can count on zero hands the number of times I’ve said “Oh great, I’m sure glad this UI is animated!” - and likewise zero times have I missed it when animation isn’t used. Animation is a way to light small units of your users’ precious time on fire, for zero benefit.
As the other user alluded to, Animations are not actually there for people who are comfortable using a computer. The vast majority of users are borderlines in capable of using the internet these days. Animations are supposed to be there to really help guide these users into understanding what the scary machine is doing when they click it. Can they be overused, absolutely, but i think have an accordion fold out animated is a reasonable case. You gotta remember your average user isnt paying any fucking attention, so drawing their attention to important changes on screen is not only good but necessary. I'd prefer no animations ever, but i also dont own an iphone while the majority of the world either does or wants to.
That's the positive interpretation, but none of the discussions I've had with UI designers or managers have been about adding animation for accessibility, and the zeitgeist of the last decade has been that skeuomorphism (of which intuitive animations are a subset) is passé.
So far as I can tell, all that the stakeholders want from the UI, animations included, is pizzazz.
Animations are also a way to explain causal relationships between interactions and their results, and to help build mental models of software behaviour.
Being related to neither software behavior nor the structure of the underlying problem, animations tend to obscure the causal relationships and make it harder for user to build a correct mental model.
Does it need animating ?
> Animating the details element is tricky. By the spec, browsers don’t natively support transitions between display: none and display: block.
Very hot take; then don't animate them!
Animation in a UI is great - you draw the user's attention to a widget that changed because they might not necessarily notice it otherwise. This improves the UX.
With a details/summary, the animation is not needed and can only make a negative change to the UX. There is no positive change to the UX that animating the details/summary elements would bring. When it is opened it is obvious.
If you really really need to animate the details, instead of animating open/close, instead animate the summary background/text color to indicate that the element has just changed state.
Would I like easy animation of open/close? Sure. Does it improve the UX? Nope.
You can put a transition on details > summary.
> browsers don’t natively support transitions between display: none and display: block.
You say that like it's a bad thing.
Fun fact: <details> even works on github and similar sites with markdown-based input. You can post large inline logs in issues without cluttering the conversation.
Details works even when it's set display:contents too, for even more flexibility. Can't animate from open›close, yet, though. That's pretty much my last frustration with it.
I think the CSS support for that has finally landed, though it means targetting a pseudo element instead. Its been a year, so support is probably good enough you don't care if just the animation doesn't happen.
https://developer.chrome.com/blog/styling-details
Note that the transition to `auto` in that post relies on `interpolate-size` which has yet to land on Firefox or Safari, and neither have movement right now.
Do they work well for when you want to preview text? Like show the first 100 characters of a paragraph and then click to expand?
Yes. For example, on Codidact (https://codidact.com), limited HTML access is offered along with Markdown when making posts, and the details and summary tags in particular are whitelisted. I've made extensive use of this in some of my content, for example https://software.codidact.com/posts/289251/289252#answer-289... . If you have NoScript you can easily verify that the expanding sections work perfectly well without JavaScript. They can even be nested, as they are here; and the summary text can contain some other forms of markup without issue. (When crafting the post, however, I had to do some tricky things with whitespace to avoid confusing the Markdown parser.)
Of course. That's the summary part
You can't actually control the open state properly from markup (the `open` attribute only sets the default state), which is why I haven't bothered with them.
I’m not sure this is correct. The DOM class HTMLDetailsElement has the `open` property, which you can use to read/write the details element’s state. If you’re using setAttribute/getAttribute just switch to the property.
https://developer.mozilla.org/en-US/docs/Web/API/HTMLDetails...
Having to use the property on the element instance, rather than the actual HTML attribute, is exactly the kind of wrapper code I want to avoid if I'm using a built-in.
You need some JS to change an attribute as much as you need JS to change a property. What am I missing?
I hope the command attribute (https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...) will eventually support this out of the box. In the meanwhile you can write a single custom --toggle command which is generic and works with any toggleable element
What kind of control are you looking for?
`open` works just like checked for a checkbox input. You can set the initial state in HTML and CSS can respond to changes in the attribute if a user clicks it. Markup won't have programmatic control over the attribute by design, that's always done in JS by modifying the element instance.
If you specify the same name on each `details` element they behave like an accordion automatically [1], no need for JavaScript. If you set one of them to open that one will be initially open.
[1] https://developer.mozilla.org/en-US/blog/html-details-exclus...
different things with the same name? O_O
It's not about using it as an accordion, it's about controlling the open/closed state from other interactions.
Out of curiosity, why have you needed to? This has never come up for me.
For a simple example, imagine buttons that opens a modal with a particular arrangement of open/closed accordions for each button.
Last I checked that without JavaScript details / summary has accessibility issues. That is, you need JS to add aria-open or similar.
It’s odd and frustrating that such an essential tag is not defined to be accessible, afaik.
Why would you need aria attributes if it's summary tag?
I presume it has to do with AT and how it typically interprets the tag. TBH IDK other than I’ve seen it come up in accessibility discussions / groups.
You can’t force it to be always open on desktop and collapsible on mobile. That was a deal breaker for me.
The interesting part here isn’t “no JavaScript”, it’s that HTML already covers more use cases than people remember (forms, dialogs, validation, navigation).
I ran into this repeatedly while writing my book "You Don’t Need JavaScript"[0]: most JS in these cases isn’t adding capability, it’s compensating for forgotten platform features.
[0] https://theosoti.com/you-dont-need-js/
Agreed! I assume the reason for the forgetting of the features is that at least some were poorly supported when first released so developers create workarounds that then become the de facto standard.
It has been amazing to see the speed up in release and support of new CSS features over the last couple of years! Even the masonry layout has finally reached an experimental stage
Yup, at this point it feels more like habit than necessity. People learned to build things like dropdowns in JavaScript years ago, so they keep doing it that way.
A lot of devs simply don’t look any further when it comes to what HTML and CSS already provide.
Most of this is great, except for the input/datalist bits, which are not sufficiently functional to be used in any real scenario. Users expect these interfaces to be tolerant of misspellings, optional sub text under each option, mobile ux niceties, etc -- and so everyone builds this with js...
My main beef with datalist is that there's no easy way to show and allow only text (e.g. Beverly Hills), but have the actual value selected be a number (e.g. 90210). In other words there's no analogy to <option value="90210">Beverly Hills</option>.
That and there's no HTML way to interactively load results. Or are you really going to serialize half a million records to HTML and transfer it all every time the relevant block is added to a page? What if it sits in the header or footer templates?
This would be possible in XSLT, if only browsers would implement the latest spec rather than abandoning it all together.
Right, I sort of expected there to be an attribute for an url.
But then you would want to send other form values along with it which might make things more complicated than it should be?Static could be better too. When search engines first started building these auto complete dropdowns the multi word input was really the killer feature. To have something like "green toyota" you would have to generate an element for all color and brand combinations? And the you want it to work for "green toyota corolla" and you get an abc kind of list length.
Perhaps a wildcard would have been fun or regex options.
> Each <option> element should have a value attribute... It can also have a label attribute, or, missing that, some text content, which may be displayed by the browser instead of value (Firefox), or in addition to value (Chrome and Safari)... The exact content of the drop-down menu depends on the browser, but when clicked, content entered into control field will always come from the value attribute
This seems... underspecified. Not ideal that Chrome/Safari aren't aligned with Firefox here, and that there is no standard way to only display the label
[from]: https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
And styling! The default styling of datalists in most browsers is just ugly.
Just been through several frontend interviews in the last few months, where it's clear that they still judge a developer's JS skills (especially React) than being semantically correct on HTML elements.
Every question/exercise is centred around how well you know React hooks, effect, memoization, modern css-in-js etc. Given I've been working with Astro recently, in one interview I talked about DOM APIs and I can see the interviewer raise an eyebrow. In later stage, even I that passed the exercises, still didn't get the job.
Remember that a large part of hiring is finding someone who fits in an existing team. A team that uses react won't appreciate someone choosing to use native DOM APIs instead of a react component.
Eh. I build apps with Preact, but I prefer candidates who know the core web platform. They’ll be more apt to use the right tool for the job and not be baffled by edge cases.
Having a separate css file make small components so much cleaner. I am not against tailwind, but I wouldn't want to use it in front-end interviews.
Because nobody outside of the HN-sphere cares about HTML purism, nor should they.
Does it bother anyone else that it links to Codepen instead of just putting them on the page?
Like I get this is a blog system but it still feels odd, especially for a "use this plain HTML"-style post...
It doesn’t really bother me.
It seems to link to the authors codepen. If you us code pen you can bookmark the snippets. Codepen colorizes the html/css etc.
Link rot is a thing though, so it’s not always ideal to have dependencies on third party urls staying the same.
Bothered me greatly. Should have done on-page AND a link to codepen for you to play with
One thing I am quite hopeful for is customizable selects! It's in WHATWG stage 3 right now. I have seen so many horrors with javascript-based custom dropdowns components. https://developer.chrome.com/blog/a-customizable-select
Plain HTML is very cozy to me - I came of age in that era. Marquee tags 4eva.
But as much as I hate to admit it, it is very difficult to build something functional today with plain HTML and no/minimal JS. If you want, say, a model form that manages its children as well, you're basically going to end up with a 2003-era ASP-feeling application with way too many views and forms (as seen on your employer's current HR system). Or you use HTMX... and you still end up with just as many (partial) views, but now with so much implicit state that you're veering into write-only code.
I dislike modern JS to the extent that I opted for Phoenix LiveView, just so I could achieve interactivity without ever having to touch JS, but in truth it's not a comprehensive solution. Still had to write a web worker, a bridge to handle things like notifications, etc. Plus the future direction of Phoenix, all in on AI, is worrying.
Honestly, I should probably just swallow my disdain and learn to appreciate and use modern JS, as painful as that sounds. I want to write and release cool things, not get caught up in navel-gazing language wars.
> But as much as I hate to admit it, it is very difficult to build something functional today with plain HTML and no/minimal JS.
I would certainly agree that using a little JS can get you further than just HTML. But I think that a plain HTML page is far more pleasant to use (and thus, functional) than the JS monstrosities that dominate the Web today. There's a reason people use the NoScript addon: because a whole lot of website designers use JS in ways that make the experience a ton worse for the user.
> There's a reason people use the NoScript addon
To be snarky, do they? The average user doesn't even know what JS is.
Users want websites that are fast and solve their problems, with a good UI. They don't care how it's made.
Make websites that people enjoy using. A good developer can do that with any set of tools, though a no-JS approach is limited in scope.
It's not an either/or. Modern Javascript is actually really nice to write and use, and you can write it in a tight, minimal way that doesn't bloat the page or slow it down.
Of course you can, but most people still opt to pull in a whole framework (React) or heavy library (jQuery) just to achieve what's essentially a few XMLHttpRequests and some DOM changes.
>you can write it in a tight, minimal way that doesn't bloat the page or slow it down.
Yet most people don't.
There are some problems with the language itself but it's mostly from a users perspective that I find it frustrating.
> Or you use HTMX... and you still end up with just as many (partial) views, but now with so much implicit state that you're veering into write-only code.
You're overthinking htmx then. I do some fairly complex stuff with no extra partials. Trick is just always rerender and use hx-select and hx-target to slice out the bits you want to update on the current page.
Server always has authoritative state and code is dead simple to reason about.
> You're overthinking [noun]
Yes, almost certainly!
> I do some fairly complex stuff with no extra partials. Trick is just always rerender and use hx-select and hx-target to slice out the bits you want to update on the current page.
Good trick! My only experience of HTMX in production entailed porting Stimulus code, hence the partials, but your approach is obviously much neater. I'll give it a shot, next time it might be suitable.
OT: marquee tags were a missed opportunity to implement horizontal scrolling often used on shopping websites. Now it uses JS to achieve the same.
I have been trying to find other more commonly known UI patterns that could be done natively. The time has long come for tabular data to be put into HTML tables just by referencing the source. Xslt almost did that. Another one is integrating xml http requests with native html. I think HTMz came close to this.
There's a usability and design issue with that as you lose what you're reading as it scrolls off the screen. Also, scrolling is a styling issue and not a document description issue which is what HTML is for.
Note: <marquee> has never been part of any HTML standard since the beginning except the current one which only has it for the purpose of marking it obsolete so people will quit using it.
I haven't used it in anger myself, but if you know Elixir and Phoenix you might like Gleam, which compiles to Javascript.
I appreciate the helpful reply, but I think this is precisely the kind of indirection I need to avoid. I'm a sucker for elegance. If left entirely to my own devices I'd probably design a language / write a transpiler of my own, and wind up with exceedingly elegant tooling for websites, and no websites. :)
This is why I don't use Typescript or frameworks in my own projects, I just constantly seek the cleanest abstraction and never get anything done. Using a deliberately messy solution is annoying but at least I accomplish stuff.
Before anyone wastes a lot of time like I did trying to use the popover API: it is not ready yet. You can do very basic things in all browsers but positioning is still different and/or totally broken per browser.
HTML and JavaScript serve distinct purposes, making better or worse comparisons logically flawed. Complex/interactive web apps requires JavaScript, period. Attempting to build sophisticated apps solely through HTML (looking at you HTMLX) eventually hits a functional ceiling.
I dont think anyone is arguing Google Earth should be pure HTML. But it is equally false you cant do Gmail with HTML only.
There are things that HTML could do, and should be doing, that is not done or not yet possible simply due to hype and trend from browser vendors. We could continue to polish HTML + sprinkle of Javascript to its absolute maximum before hitting JS Apps. Right now this is far from the case.
Gmail used to provide an HTML version. It got removed only recently
Gmail with html only would not be a nice experience. Modern gmail is really bloated but it's actually one of the few web apps I have no problems with.
Gmail used to have an html-only version if I remember rightly. Perhaps still does. It was faster and perfectly usable.
Hey the email services has just proved you could offer better than Gmail experience with HTML + small dose of JS. Another example being the new FE on Github.
At the end of the day it isn't really the tech that is the problem. Is how people use the tech. And for thousands of different reasons keeping it simple has always provided better experience evaluated on the whole.
> Another example being the new FE on Github.
Github's old frontend was mostly HTML with a bit of JS, their new frontend is react. The old UI had its bugs, but it was much better than the react version in my experience. I still commonly find the UI out of sync with itself requiring a reload, but now I also frequently wait for the page to load and viewing large diff's is a performance nightmare.
I assume you mean htmx. It doesn't have to be either/or. You can supplement htmx with Javascript.
The core idea with htmx is that you transfer hypertext with controls and structure built in, not just a JSON blob that requires additional context to be useful.
I have just shipped a very useful and interactive app surprisingly quickly for my customer using just htmx with a little Javascript.
It shouldn't have to be this way though. There is no reason html can't do things it needs to do to build complete apps. We could use reasonable defaults to allow a new type of html markup without JavaScript.
All the http verbs. Decent html input controls What else?
Depends on how complex it is meant to be. Just like many wordpress sites that could easily have been static websites, many javascript heavy sites could have easily just have been using htmlx.
If your need really, goes beyond what htmx offers, then you may need Javascript. But in my experience people tend to use the tools they know for their job, not the tools that would be best suited.
FYI, it's easy to cache the html output of a WordPress site, resulting in essentially a static site with graphical admin, page builder, and all the other bells and whistles.
I don't want to be so negative, but we got details and popover after literally decades, and we still have datalist presented as a plausible option? Html is so underdeveloped and the first we all agree on that the first we'll pretend from committee real advancement
i like the points the article makes, but i really wish it used looping videos instead of actual GIFs
i don't really see any reason to use GIFs here; any widely available video codec like H.264, VP8/VP9 or AV1 will result in significantly smaller file sizes, look better, and will allow enabling controls for seeking and play/pause
Progressive enhancement is the way to go if you care about technical excellence. For some reason it fell out of fashion.
https://en.wikipedia.org/wiki/Progressive_enhancement
> if you care about technical excellence
Or accessibility.
It fell out of fashion because the entire web was consumed by react.
> Input with Autofilter Suggestions Dropdown
It's great until you have a typo in the field, or want to show options that don't start with what you typed in but appear near the end of an option (think Google search's autocomplete). There's no way to filter in Javascript and force it to show certain options with <datalist>. I've resorted to <ol> for search suggestions.
I don't see any mention of HTTP 204 or multipart/x-mixed-replace. Those are both very helpful for implementing rich JavaScript-free HTML applications with advanced interactivity.
When building out a new app or site, start with the simplest solution like the html-only autofilters first, then add complex behavior later.
It's good to know these things exist so there are alternatives to reaching for a fat react component as the first step.
Until your client tells you that it doesn't work in Edge and you find out it's because every browser has its own styling and they are impossible to change enough to get the really long options to show up correctly.
Then you're stuck with a bugfix's allotment of time to implement an accessible, correctly themed combo box that you should have reached for in the first place, just like what you had to do last week with the native date pickers.
Right, don't add complexity until you have to.
Have you no sense of craftsmanship?
It’s great to see practical examples that push us to consider what the platform already offers before adding more layers of complexity.
The Pentagram at the top of the page does not load without JS enabled.
So it's.. fitting, how meta.
I’ve tried replacing my modal components with the <dialog> element, but had to reverse everything due to this issue: https://github.com/whatwg/html/issues/9936
In short: you can’t have an interactive popover (e.g. a toast notification) on top of a dialog modal.
I’d love to use the new native elements but we’re sadly not quite there yet.
I didn't know about <datalist>, but how are you supposed to use it with a non-trivial amount of items in the list? I don't see how this can be a replacement for javascript/XHR based autocomplete.
Don't use it, it totally blows. For another oddity to not use, check out the multiple select: https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...
Expecting users to press modifiers when clicking on these is so funny.
> Expecting users to press modifiers when clicking on these is so funny.
I mean… 5 year olds can figure out shift-click in Minecraft.
Trying to figure things out in a game is fun! Trying to figure things out on a website is a sign the UI sucks.
> I don't see how this can be a replacement for javascript/XHR based autocomplete.
It can't do complex autocomplete. It's ok for simple cases only. I use it with a 25k long list to ease the input. Works well enough for this.
> If we can hand-off any JS functionality to native HTML or CSS, then users can download less stuff, and the remaining JS can pay attention to more important tasks that HTML and CSS can't handle (yet).
You can't. It's only supposed to be used for a limited list.
And even if you allow XHR and add options to a <datalist>, it still has terrible UX.
Brilliant I will be adopting a few of these, I have been on a personal quest to reduce js use recently, I feel like I spend more time debugging js than producing the end result.
Nobody tell them about how much stuff can go into SVG. That can even be inlined within HTML source code.
And you can target it with css! Slap a class attribute on a path element; it’s fine!
I love semantic html. It feels so much more modern and efficient than all of these fashion-driven front end tools.
It seems like an oversight that there isn't a native way for a popover to be activated by hovering over the target element.
I was trying to rewrite some UI library with html sometime ago following the W3C accessibility specs and found out a lot of patterns can’t be done with pure html and require javascript unfortunately.
like what?
Tabs, accordion, combobox. There is a whole lot more, these are just the ones I can remember now.
If we've concluded that's it's okay to have elements that change/morph, as we seem to with the introduction of things like details, a native tab-like element feels like a glaring omission. Tabs have been a long-standing UI pattern and forcing every site to implement their own is a nightmare for accessibility. (The page you're reading is maybe already in a browser tab.)
I wouldn't be surprised if it turned out less than half of the custom tab interfaces on the web failed from an accessibility standpoint. When considering ARIA guidance, I don't even think it's possible to build an accessible version in HTML alone.
Other people have recognized it's missing. Open UI has a draft spec for it[0] and CSS Tricks has an article from 2001 about Open UI's experiments with sections for tabs[1]. I have no idea what happened on this front, though.
[0] https://open-ui.org/components/tabs/
[1] https://css-tricks.com/newsletter/281-tabs-and-spicy-drama/
Accordion behavior is discussed in the article in the "Accordions / Expanding Content Panels" section:
> Use the same name attribute on all related details (like radio buttons) to restrict only one open panel at a time
And tabs can be a <details>-based accordion with some creative CSS to adjust the layout (left as an exercise for the reader, but I could write up an example if that would be helpful!)
It won’t have the necessary keyboard shortcuts.
Yeah this is true at this point. A lot of more complex patterns require JS to be accessible to screen readers.
We still should do more with HTML and CSS! And reach for leaner solutions than React everywhere.
But be careful going for a pure CSS solution for things like tabs if you don’t understand the accessibility requirements.
(I wish the HTML spec would move faster on these common patterns!)
https://www.w3.org/WAI/ARIA/apg/patterns/ is my go to for accessibility requirements of components.
And yes, being able to do all of these in pure HTML/CSS would be awesome. Though we are getting there with things like `details` and the newer `popover` features which should make things like rich tooltips, menu buttons, etc. a lot easier to implement. IIRC, there are also several anchor CSS properties to make positioning a lot simpler.
> We still should do more with HTML and CSS! And reach for leaner solutions than React everywhere.
It's pretty difficult for anyone to completely understand all the nuances in HTML and CSS. It's a big mess that gets bigger and messier every year.
We should have just given JavaScript even more power over controlling the viewport and leave HTML and CSS for the history books.
All of those can be done with pure html/css, eg. https://codepen.io/mikestreety/pen/yVNNNm
Gimme a dark/light mode switch. CSS is allowed.
Use a checkbox, d. Define vars for light mode. Override when checked for dark mode with body:has(#d:checked) and can include the dark mode media query too
See https://lyra.horse/blog/2025/08/you-dont-need-js/, via https://news.ycombinator.com/item?id=45056878 (which discussion contains some deeper explanation).
Why would you build a switch instead of relying on the user’s system settings? The only reason I can imagine is that your dark/light mode is not usable/readable so it forces the user to switch
I often use different light/dark settings between apps and my system. Just because I want system UIs to be dark, for example, doesn't mean I want to read long pages of white-on-black prose on your blog.
Seconded. Just because I like to have the browser toolbar dark and GitHub dark doesn't mean I also want to read lengthy articles (LWN) in thin white text on a black background.
Having it default to the users preference is nice, but you should still provide an override. I sometimes use my browser in light mode while my OS is dark mode. Many times, I find the contrast for dark mode websites too low unless I’m in a totally dark room.
Because I don't want to toggle my whole system theme based on one special website.
Checkbox and :checked are your friends.
Your blogs have very small amount solution, but the JS use cases are very large. How this little replacement can do more thing? I usually like the idea of being using as lean as possible, if it's can be possible to do more thing just with HTML and CSS that's obviously cool. Is it really possible to replace JS with HTML in near future?
BTW the toggle solution (expanding content) is good.
This is helpful when development revolves around static environment.
The Popover API looks really cool. Could see it for tooltips or lightboxes.
A pleasant surprise to see Aaron's post here, we worked together for a bit on frontend optimization in a multi-tenant international ecommerce platform. That work was a large part of my inspiration for building https://contentblocksjs.com which encapsulated a lot of the JS concerns into web components.
I am reading the CodePen example for summary/details. Especially the CSS part.
Its so easy, like a breeze!
JavaScript is the primary language of the web, HTML is just the payload carrier.
I don't want it to be this way, but HTML has nothing approaching even one tenth the ambition of XSLT and XForms.
If HTML were proposed today, you would be laughed out of the room.
It feels like some variation of this post gets submitted here every week.
That’s a bit of evergreen topic. “stop bloating web with js” comes up fairly often and there are those people who think they found a solution and everyone should start using whatever they imagine is “best for everyone”.
In my opinion most of those people struggle with whatever they encountered in ecosystem and just want to find a way that fits them - while also trying to make others do the same.
*“You didn’t want to make things perfect. You just hated things the way they are.”*
And it will continue until node_modules shrinks!
Something I keep thinking about when I consider the trade-offs between building a site with HTML/CSS wherever possible vs JS is what the actual _experience_ of writing and maintaining HTML/CSS is vs JS. JS gets knocked around a bunch compared to "real" languages (although less so in recent years), but at the end of the day, it's a programming language. You can write a loop in it.
Writing a web server in C++ is a way to get excellent performance. So why don't most people do it?
> Writing a web server in C++ is a way to get excellent performance. So why don't most people do it?
Because they already wrote it in C.
Apache and Nginx are both written in C. Together they run 57.7% of all web servers:
https://w3techs.com/technologies/overview/web_server
JS got popular because some devs were trying to realize a world where the same code can be shared on the front and backend. I think on the surface, it was a noble goal with good intentions. Having only one programming language to handle is going to create some efficiency gains when you work in a large company with thousands of devs who all need training want to share knowledge as a larger organization.
In the past decade, we went full JS as an industry and now we’re starting to swing back. Server side interactivity like Phoenix Liveview, C# Blazor, HTMX, PHP/Laravel Livewire, Rails Hotwire, all of these are different abstractions around JS to make interoperability between the frontend and backend more manageable and they’ve come a long way to closing the gap. Advancements in HTML/CSS standards also deserve credit for closing the gap but we’re still not quite there yet.
But at the end of the day, the web is dynamic. As new tools and techniques are discovered, the industry will continue to evolve and certain “hacks” will become new standards and ignorant newcomers will reinvent the wheel again to achieve some crazy interactive design because they didn’t know any better! And it wil work, mostly.
Until the way we interact with browsers changes, I feel that we’ll continue to bolt on new features over time and the web will continue to evolve. Just like the iPhone, a surge of use of smart glasses could change the computing paradigm or perhaps its some other device entirely.
So you can (and should!) try to optimize for today, but trying to optimize for tomorrow will always carry the risk irrelevance if the market pivots quickly. Bleeding edge is risky but so is falling behind.
What would make this super easy to adopt?
A simple page which shows what native components are available with/without the need for polyfills for a given browserslist config
Some of these new HTML features don't fully work in my "ancient" browser. But all of them partially work (ie opening the accordion element doesn't close others but it still opens and closes) and they still remain functional elements I can read and interact with. This puts them far ahead of any javascript implementation which almost universally fail to nothing.
What "ancient" browser are you using?
Probably something 300 CVEs ago.
<blink>
I really try not using JavaScript unless absolutely needed. On my latest project, the whole site actually functions without JavaScript and is server side rendered. However, there's some small piece which I really needed JavaScript for couple reasons.
Basically, I have a site which collects the top STEAMD posts from places like HN, lobsters, tildes, slashdot, bear, reddit etc and displays them in chronological order. I wanted a way for users to block posts with certain keywords or from specific domains. I didn't want to do this server side for both performance reasons plus privacy reasons. I didn't want users to need signing up or something to block. I also didn't want to collect block lists for privacy reasons. So, I resorted to using JavaScript and local storage. All posts within the filter for the date are sent and JavaScript is used to block posts with keywords before displaying. So my server never knows what keywords are blocked.
Site for anyone curious:
https://limereader.com/
I'm so not impressed by the toggle implementation... How nice it could have been.
Nesting the elements is a truly hideous choice. The summary is part of the details?? I thought they were opposites.
Should we also put the headings in the <p> from now on?
Identifying a target should be done by id or by name. That it does use a name because js can't target it without makes it even more stupid.
We already had labels for form fields. Inventing a completely different method for something very similar is a dumb idea. The old checkbox hack is more flexible and less ugly for some implementations.
Why force the hidden content to be below or above the toggle? We aren't gaining anything with this.
What is this nonsense for an element to not just be hidden or displayed but to have some weird 3rd state where only one of its children is shown?
How should styling it even work for this new state? If I apply a style to the hidden content it must also apply to the link? The text is hidden but the style is visible??? Preposterous!
Don't try style <details> to avoid unexpected behavior. Try wrapping the hidden content in a new element to make it behave normally.
What is this ugly arrow? If you find 1000 websites using a toggle I doubt there is one using an ugly arrow like that.
The default styling gives no clue about it being clickable?
The pointer (awkwardly called the cursor) choice is the text selection?????
Blue underlined "more" is what everyone does and everyone is used to. The cursor should be pointer. (This is css speak for "the pointer should be a hand")
The number of js toggles you can find online where the button lives inside the hidden text is guaranteed to be zero. Forget about drop in replacement, you will have to reinvent your css.
Maybe I'm dense but I also want my url to reflect the state of the page. I would have been impressed if that was supported. Personally I use actual links and disable default action in the listener if js is enabled/working or modify the state on the server if js isn't available/working.
It would have been great if the toggle action was implemented as a simple attribute something like toggle="element name" so that anything can be clickable and anything can be toggleable. Have a "closed" as well as an "open" attribute for the target.
Doesn't seem very hard. An open/closed attribute would be useful for other things too. Using display:none is terrible as display: is used for many things.
> Nesting the elements is a truly hideous choice. The summary is part of the details?? I thought they were opposites.
It gives them a semantic connection. Last I checked, HTML isn't really based on giving special meaning to combinations of sibling tags. A summary is part of the thing that conceptually requires detailing.
> If you find 1000 websites using a toggle I doubt there is one using an ugly arrow like that.
I think the default looks fine. But TFA clearly explains right there that it can be styled. (Specifically, by styling ::before on the summary tag.)
> The default styling gives no clue about it being clickable?
You asked what the arrow is, and then asked about the lack of indication that the summary header is clickable. The arrow is exactly that indication.
> Maybe I'm dense but I also want my url to reflect the state of the page.
If you scroll, should the fragment automatically update as you scroll past anchors? I think I'd find that quite annoying.
The problem is that it's difficukt to style or animate those things. Unless you're builsing something for dun or technical where it's not important it's fine but i doubt any real world commercial project would be satisfied with just this