My wife and I have watched a lot of TV in the past year with the volume down and captioning on, while we enjoy some down time while our baby is sleeping.
A ramped entrance to a building allows wheelchair-bound folks access, but it also helps able-bodied people using delivery dollies.
Making simple, lightweight web pages makes them accessible to those with older browsers and devices, but it also saves power consumption, time, and frustration for folks on M1 MacBook Pros.
I really love how Microsoft's design team has pushed this with their “Inclusive Design” concept[1] and highlighting how any sort of disability can be situational (holding a baby), temporary (broken arm), or permanent (missing arm) but we often only think of those in the last context despite there being orders of magnitude more people in the first two categories. Parenting is really good for showing this out since you get the need to be quiet, be able to do something using only one hand (or none), navigate sidewalks with a stroller, trying to do anything while sleep deprived, etc.
There's a great chart on page 42 (“Persona Spectrum”) of this PDF which I try to get every team I work on to go through which really gets people to think about what this means:
https://download.microsoft.com/download/b/0/d/b0d4bf87-09ce-...
2. Microsoft makes software which costs a lot of money (for most people in the world); often a lot of money. One of the main problems people with disabilities have is low income and lots of expenses.
3. Microsoft makes software which hogs resources and thus runs poorly, or not at all, on older or simpler, low-power devices. This also not very accessible. So it might have a perfect UI experience - on the computer people don't have. Admittedly, though, this is not as bad of a problem in the past few years.
4. I wonder how actually accessible MS apps and Windows actually are. Considering how they've botched a lot of their designs in recent years even for normies (Metro; Ribbons).
Similarly, while Microsoft software isn't free nothing in computing is and it seems like an odd angle to criticize an effort to help make their products better based on decisions made by other people in different parts of a huge company, or to criticize them for things in the past which they've since corrected.
Luckily, that is not the case:
https://en.wikipedia.org/wiki/Free_software
I suggest you take a bit of time to read some Free Software philosophy. Consider starting with "Why software should be free", by Richard Stallman:
That free software has to run on non-free hardware (even if the oft-heralded RISC-V revolution happens, someone has to manufacture the device) and most people do not have the skills or time to assemble the hardware or maintain an OS. That means that we're all making tradeoffs of what things we buy and what we do ourselves, and for the vast majority of computer users, even open source developers, that involves outsourcing that work to a handful of companies (or, maybe, a hardware company and Debian). Most computer users consider that a good deal: they spend time on things they like and pay a rather small amount of money relative to the utility they get from a transformative device like a general purpose computer. This means that for most people the decisions which matter the most are made by Microsoft, Apple, or Google — and that anyone who cares should applaud them improving accessibility because it will affect daily life for billions of people.
This is especially true for accessibility, where people develop deep habits around things like screen readers and quality matters a great deal. I happen to work with a number of visually impaired users and have not heard good things about the Linux screen readers compared to their Windows or iOS counterparts, and that means that in the context of this thread it's really not useful to snark about the moral superiority of free software.
> open source software is only free if your time and expertise are free
but since you seem to actually be trying to discuss this in a thoughtful way, please note that we're not given a choice between free software that takes time and expertise to use and maintain vs closed software that takes no time or expertise to maintain. To mimic your original phrase, Windows 10 Pro is only $199 if your time and expertise is free. Going further, in my experience it's often been the case that the free software takes less time and less expertise to use and maintain.
Also, it's worth mentioning that you are conflating two different things, open source software and free software, a distinction that often doesn't matter but is central when the point at hand is the ethics of free-as-in-freedom software vs non-free software, a distinction the term open source was deliberately created to elide. You are also conflating free-as-in-freedom with free-as-in-beer by opposing the "software is only freedom-free" with "your time and expertise are free-as-in-beer free".
Finally it's not clear to me why you went on to address the pragmatics of non-free hardware or the fact that human effort is necessary to build computers, maintain distros, write kernels etc. Is there a claim that it's unethical to try to run free software on the hardware that you have, or that free(dom) software must always be provide without cost on hardware that is both free(dom) and without cost? If not, then computer users can still spend a small amount of money, relative to utility, to buy a computer that runs free(dom) software.
I'm not disputing your claims that accessibility software on Windows or iOS is better than that on Linux, because I don't know the space. It's an unrelated argument, afaict. To illustrate this, simply imagine that some government or corporation had decided to make high-quality accessibility software available under a free license. (I am reminded of Intel's work supporting Dr. Hawking.) You wouldn't conclude from this that all your previous thoughts about free and open source software vs closed/commerical software were wrong, I assume.
I just wanted to point out the irony that the manual and other documentation for this "inclusive design" toolkit come as PDFs – the worst choice you can make if you strive for accessibility.
https://en.wikipedia.org/wiki/PDF/A
PDF/A is a small family of sub-formats to PDFs which amount to the accessible subset of PDF + some requirements.
PDF/A and PDF/UA are mandated by a variety of governments for PDF distribution. They're very suitable to archival and are accessible (screenreaders can work with them, their text can be cleanly copied, etc).
HTML, on the other side, can be easily made readable on any device. Even if the author didn't care, through readable mode in my browser.
Why would you read PDFs on a smart phone anyway?
Is there another option here?
IME PDF/A is better than nothing, but far worse than plain HTML for accessibility..
If it's scanned copies then all it is is a bunch of images, not A compliant.
As long as it has actual text and is reasonably typeset (selecting text in this pdf seems to work properly), I don't see what's wrong. I'd take it any day over webpages with random floating content that you have to click through and use a modern browser with javascript without blockers/filters for the content to even render, in small text on 20% of the page, with low contrast and somehow broken zoom.
Where it breaks down is some badly designed OCR systems which place the letters independently rather than in lines/blocks. But that's starting from a document which was not accessible in the first place, so not a huge issue.
I have to admit that many modern SPAs also won't rendet on many devices, though.
This really challenges the view that thinking about disabled users is catering to the needs of a small group of the population. Whereas in fact it is bringing benefits to the majority of the population (at some point in their lives).
Disability is just the fine print.
I remember walking around in a city after a parent died, and the world seemed like an illusion.
Then I wonder if in a parallel universe I'm already dead.
Then I realize that if there are parallel universes there are probable countless times where I'm already dead.
Ever used your phone in the blazing sun, or rain? Had a partly shattered screen or tried to do something on the stone-age library computer? Searched for something with really slow or shaky connection? Tried to explain where to click via phone?
Suddenly all those accessibility features and "fail gracefully" come in really handy.
"We're all just temporarily abled."
is quite famous quote, as far as I traced attributed to Cindy Li.I like how Apple TV now lets you pair two bluetooth headsets. My wife and I can watch shows without waking the baby. Combine that with the visual sound indicator on the baby monitor, and headphones with transparency mode, we can see when the baby is making noise and hear other environmental noises too. It's a nice combination of accessibility features that make a clever setup for hearing the TV. I can't wait until Apple TV supports spatial audio... then we'll have "surround sound" in our headphones too!
I use closed caption when possible because often I don't quite hear things right and can't make out words. I could turn up the TV, but this means I can enjoy it at the same volume as everyone.
There is also sort of "exclusive design" that is practiced. You know - where you can't get comfortable on a park bench because you might sleep on it, or a ramp or stairs are built to exclude skateboards.
Same thing happens with websites, to "control" adblockers or "reader mode" and other nonsense.
Did you mean to say ". . . for folks /not/ on M1 Macbook Pros"?
This seems to imply that web sites are businesses. If my hobby sites are subject to this, I'm going to be in trouble.
(Complete list of public accommodation is 24 USC §12181(7)--https://www.law.cornell.edu/uscode/text/42/12181).
> The term “commerce” means travel, trade, traffic, commerce, transportation, or communication
It is pretty funny that their definition of commerce includes commerce, but seems like communication is on the list as well.
> The expression "platinum" means platinum, iridium, osmium, palladium, rhodium and ruthenium
https://ec.europa.eu/digital-single-market/en/web-accessibil...
How big an issue this is depends largely on how big your hobby site is.
Update: I think this might even be wrong, its likely personal sites are fine. It's still good practice though!
What was the catalyst behind this trend? I just don’t understand the reason behind it. Hasn’t <button> been around forever? Was there some limitation on the button tag that made the div tag more desirable?
There’s no good reason to just not use buttons these days, but for a long time that wasn’t the case (cross platform).
And, as you mention, the default styling was typically "make it look like the button in the underlying OS".
So there was a lot of inertia from existing code, examples, and docs that predated buttons. And a (misinformed) avoidence of it when it did appear because of styling.
Since it's introduction, jQuery was born and died. Backbone was created and became obsolete. React took over the internet. Webpack, babel, and node were invented.
With all that change, I think it's safe to say web technologies are adopted or not based on inertia.
I think you are being over-generous.
On the other hand, the Compose "button" in gmail is ... a <div>:
<div class="T-I T-I-KE L3" style="user-select: none" role="button" tabindex="0" jscontroller="eIu7Db" jsaction="click:dlrqf; clickmod:dlrqf" jslog="20510; u014N:cOuCgd,Kr2w4b" gh="cm">Compose</div>
FWIW, my issue with using divs is not that they should never be used as buttons. It's that few people who use them that way go to the effort of doing it correctly. It's absolutely possible to do it right and have a functioning/ accessible js button.
Either go the full way and do it right—with accessibility—or just style the standard components. In most cases the latter is the easiest way to go if you care about getting it right.
Also, FWIW some of the problem is with the browser makers (still!!). CSS still has a lot of browser specific quirks when working with some components.
Designers absolutely need to be aware of and work within their platform's constraints, like a painter needs to understand and work with the physical characteristics of their canvas. It's this kind of mentality that results in visual specs that read "Make the window look exactly, to the pixel, like this Photoshop image!"
This subthread is about why divs are used for buttons and why developers reimplement existing HTML components poorly. Good designers help their teams avoid this, good engineers work with their designers to avoid it.
0: https://www.gov.uk/government/organisations/government-digit...
It’s funny - I’d expect an interior designer to know something about the limitations of the materials they work with.
Why don’t we expect this of digital designers?
I think that's somewhat like what the relationship between designers and developers should be.
It looked good. But hardly worked.
Apparently the web hasn't really changed in that part: we're still churning out HTML that hardly works, but "looks good".
"button are special element and their content is always vertically aligned by default. Even all the propertise are the same, if you don't center the text in the link, you won't get the same visual"
https://stackoverflow.com/questions/64766922/what-can-make-t...
I created a minimalist CSS framework called Neat CSS (https://neat.joeldare.com). I use anchors instead of buttons and explain why in the quote below. I'd love feedback on the accessibility of that alternative.
It's best to use semantic web tags whenever possible. Buttons are a unique case where you're typically linking somewhere, but the button tag doesn't currently support the href attribute. So, buttons are anchor tags with a class of button.
An anchor tag that links to another page but looks like a button is probably fine. So long as you can access them via the keyboard (tabbing to it) and they have a distinct appearance when focused so if you are tabbing through you can see whats active (this is different from the hover attribute).
You can drag it to your tab bar, bookmarks or desktop, middle click to open in a new tab, copy the target URL etc.
With a button you can push it.
> This is a fantastic article, it does miss one very key point about bog-standard HTML which is worth mentioning though.
> The standard widgets are all accessible, people with screen readers, limited mobility, poor vision, etc, all rely on pages being written so the devices they use to read the web can function properly. People who choose to eschew these standard components very frequently end up with a site which is unusable for disabled people.
> Very few people have the skill to turn a div into an effective button, yet over and over I see web sites with rows of buttons which are in fact just divs with javascript behind them.
Using a div as a button is not "bog-standard" HTML or a "standard component". Use a button and be done with it.
Using React or whatever does not prevent you from doing this, and in fact makes it less obvious to anyone doing a cursory review. I've not seen any difference overall in the amount of people knowingly using divs as buttons with plain HTML vs whatever framework, but I have seen an increase in unknowingly doing so by importing a calendar widget or similar without reviewing it.
Fully agree. In fact that is exactly point I was trying to make.
<div className="button" onClick={handler}>
and it immedaitely sticks out as a mistake. Because its easier to statically understand these mistakes, theres even tooling to automate detecting some of them as you type https://www.npmjs.com/package/eslint-plugin-jsx-a11y> <div className="button" onClick={handler}>
> and it immedaitely sticks out as a mistake. Because its easier to statically understand these mistakes, theres even tooling to automate detecting some of them as you type https://www.npmjs.com/package/eslint-plugin-jsx-a11y
I'm not sure how you got that out of
> I have seen an increase in unknowingly doing so by importing a calendar widget or similar without reviewing it.
I'm referring to:
<div class="div-button" onclick="clickHandler()" />
VS: <SomeGreatNpmWidget onClick={clickHandler} />
The point of using components is so that you don't have to look under the covers to understand them at a high level, but that also means if you don't look you don't see how they're breaking accessibility.> Using React or whatever does not prevent you from doing this, and in fact makes it less obvious to anyone doing a cursory review
Sure. From that level it might not be apparent, but within that component the mistake would be immediately obvious, on a cursory glance. Again, I don’t think React or whatever makes this less obvious.
But really, this just isn’t a problem I’ve seen in practice. Because of good tooling, and the ease of doing things the right way, I think the majority do it well. Those that don’t were going to fuck it up regardless of the framework or paradigm.
You can add role="button" and a modern browser will treat it just like a normal button and all the accessibility that comes with it.
[1] https://darekkay.github.io/presentations/accessible-web/reso...
HTML covers text, painfully basic tables, semantic rectangles (<nav>, <aside>, etc.), form inputs, and embedded media. Basically, anything you could build with a word processor. Outside of that, you're on your own.
JavaScript and CSS are continually improving and are so much more capable than they were even five years ago, but it feels like HTML is just stagnating. Why can't we have new elements every year, with behaviour? Stuff that works just fine out of the box, but can be easily tweaked with CSS or a little sprinkling of JS. The amount of developer time and bandwidth we could save not having to re-invent the wheel for every little thing.
but your cited site is not a good advocacy of that. Not even the most exalted UI framework fix the most obvious problems with those either. e.g. tables still won't hold the header and first column(s) while scrolling.
Most of the these elements does not help consuming information, like carousels. And any true UX designer abhor them all for the fallacy they are.
https://caniuse.com/css-sticky
position:sticky works on tables.
even the grand-parent, javascript infused solutions, only two out of 20 support this extremely basic and obvious use case.
Web UI standards are a joke. Everyone involved only cares about rich ads and accordions/carrousels which are lame ways to shove lots of content in a badly designed space.
[1] https://addons.mozilla.org/en-US/firefox/addon/axe-devtools/
[2] https://chrome.google.com/webstore/detail/axe-web-accessibil...
1. https://github.com/chanzuckerberg/axe-storybook-testing 2. https://storybook.js.org/docs/react/get-started/whats-a-stor...
edit: actually it's spelled https://pa11y.org
Products like Siteimprove are targeted at Enterprise (e.g. demos with sales weasels, unpublished pricing, etc). They mostly suck but check an auditing and compliance box.
It might not get rendered the same everywhere depending on the device receiving it.
If the goal was to send pixel perfect pages, the web would have settled on postscript + hyperlinks.
> Courts have rules that the ADA applies to web sites along with any other type of business
For a business like a coffee shop I'd imagine that accessibility is codified in law through things like building regulations (ie, you must have n disabled parking spots per square foot, a disabled toilet, etc)
Is there an equivalent legally codified and enforceable standard for the web?
More or less yes. The link I posted has a few good resources, but generally speaking sites need to work with a screen reader and they need to be accessible via the keyboard (which by extension enables other accessibility devices).
This site has some more guidelines. https://www.essentialaccessibility.com/blog/ada-guidelines
Generally speaking, if you start with existing components that do the thing you want done, you have a huge leg up. If you try to reinvent the wheel, you are likely in for a lot of trouble. The more complex the control, the harder it is to roll your own. Calendar controls and Selects are particularly tough to get right.
UPDATE: Also, the wc3 has some good guidelines which should cover your ass.
> In this course you’ll get hands-on experience making web applications accessible. You’ll understand when and why users need accessibility. Then you’ll dive into the "how": making a page work properly with screen readers, and managing input focus (e.g. the highlight you see when tabbing through a form.) You’ll understand what "semantics" and "semantic markup" mean for web pages and add ARIA markup to enable navigating the interface with a range of assistive devices. Finally, you’ll learn styling techniques that help users with partial vision navigate your pages easily and reliably.
Gahhh. I use either an <a> or a <button> and put style on it. Has all the same visual appeal of the equivalent div. But it already is a button.
Generally, it should be noted that an <a> tag is only appropriate if it makes sense to right-click this element and get options to "open in new tab" or "add to bookmarks". All the stupid "buttons" implemented as <a href="#"> + e.preventDefault() need to die.
Later on, when people started treating web design as a general purpose graphical and GUI framework, many of the accessibility benefits probably vanished.
General accessibility be damned.
<noscript> fallback messages be damned. (Except when snitching on your page views to Google Analytics, that's too important to be neglected! I wonder if GA has any default marker to let their customers distinguish these views from normal ones, or if that's still left to each customer to implement manually.)
UI customizations in the form of browser extensions and styles be damned.
Ad blocking be damned.
The standard widgets are all accessible, people with screen readers, limited mobility, poor vision, etc, all rely on pages being written so the devices they use to read the web can function properly. People who choose to eschew these standard components very frequently end up with a site which is unusable for disabled people.
Very few people have the skill to turn a div into an effective button, yet over and over I see web sites with rows of buttons which are in fact just divs with javascript behind them.
Making your site accessible isn't just a good practice, in the US, it is the law. Courts have ruled that the ADA applies to web sites along with any other type of business. A small sample of sites where people thought they could smash together super-fancy sites without concern for basic needs of users: https://www.atilus.com/top-10-ada-lawsuits/