Hacker News Reader

Comment by falcolas | Hacker News Reader

What bothers me more than people designing inaccessable sites is their excuse of "well, why put extra money and effort into catering to such a small segment of the population?"

Videos released without captions are OK, because "they're free!" Websites without Aria tags are OK, because "screen readers just need to get better." Low contrast fonts are OK, because "my design won't work with different font colors."

Everyone who doesn't die in their prime of life will eventually be disabled in some fashion; but so few people really seem to realize this (or care).

EDIT: Good question - do I practice what I preach? Yes, I do. I'm in an SRE position, and a few of our developers are hard of hearing, so I caption the training videos I make, without asking if they need them. No blind folks, but I make sure to use reasonable styles and page layouts to support fading eyesight. I don't set up Aria tags myself, because I don't have any users who need them at the moment.

None of my external or OS contributions are to UI related projects, so nothing of note there.

View on HN

Your point about profitability assessment is an excellent one. I recently spoke to the CEO of a company that makes ebook reader software. This CEO has issued multiple open letters about accessibility, claiming that his company cares a lot about making its products accessible.

But in a conversation with him about a new text accessibility tool, he said that if he can't "pass on the cost to his customers", he is unlikely to offer it to his customers.

This shows a complete misunderstanding of what accessibility is. If his company could make more money by including a new feature, then integrating it wouldn't be "accessibility" – it would be "usability" or "good design".

Accessibility is (almost), by definition, the stuff you do to make your product usable even though the time you are spending does not pass a typical cost-benefit test.

And the point of ADA lawsuits is to make sure that the cost-benefit does line up, if legal fees and judgments are included.

The truth is there is a cost associated with all of this. Example my current job I brought up issues of font scaling, screen readers, etc. They care to make it good enough BUT not enough to invest significant time, because we need MVP. People won't accommodate the minority unless they are one, unless they have to, or unless they really really care and can convince their boss. This is an unfortunate truth of humanity or business.

Personally I very much care. But I don't get to call the shots.

I was going to make this comment on the parent comment but you really tee'ed it up for me perfectly.

Not doing basic accessibility on the web is just mental laziness, plain and simple. I have to strongly disagree the the implied point in your post that it requires investment of any significant time.

You can get 90% of the way to "good" just by picking appropriate font sizes and colors, doing halfway sane IA, and probably most important just using semantic markup. Adding text info to images, aria tags, etc are very minimal efforts if you simply have it in mind as you work.

None of those things require any extra time except that people need to care enough to read some documentation and watch a couple youtube videos which is a one-time up front investment of 1hr or so (not saying you are an expert after 1hr but you can definitely do "ok" with 1hr of research and learn as you go from there). a11y gets more complicated when you start doing heavy js with custom realtime interactions but even then you can get 70% of the way there with very little effort and it is a bit of a red herring because how many people are really working on fancy modern SPAs in their day jobs?

> I have to strongly disagree the the implied point in your post that it requires investment of any significant time.

Totally agree. Basics like alt tags on images, wcag contrast friendly colors and semantic tags to let screen readers know this is a repeated menu, this is content, etc... take very little additional time and help so many people. Even in SPAs these items are basic.

"this is content" sounds extremely useful for everyone. Curious if there is much abuse of accessibility markup to trick users into clicking ads.

There's an argument to be made that the aria data can be leveraged by your test automation as well.

Code that a screen reader can't decipher is pretty hard for test frameworks to figure out, too.

Hey, ever since we started using REMs it made scaling much easier as well. There are times when we had to work around for example bugs in Chrome with font scaling and transform under certain conditions which made chrome resize things incorrectly, but only chrome.

Its little things that take up a bunch of time. In theory we're already 80% of the way there, but the last bit is what makes the big difference.

Black text costs more than grey text? What, are you paying for ink?

Sometimes I have a hunch that inaccessibility is what costs money. A basic informative HTML web page is cheap to write and host. The deluge of swirling graphics, icons, and microscopic pull-downs, are what cost money to create.

Yes, if you make a mess and bring in some accessibility compliance people that have to suggest all kinds of shitty workarounds it will cost you money. If your devs don't know why those shitty workarounds are there they will break it all again in a next release. If you just use semantic HTML from the start you'll have a more robust product and less maintenance in the long run.

Oh, and ARIA rule one: don't use ARIA, use the native HTML equivalent. Rule two: Only use ARIA if you can't express your intent in native HTML and make sure you know what your code is doing, don't just copy paste some ARIA example from another project or random site, lots of ARIA in the wild is wrong.

I'm a blind developer and accessibility consultant. Accessibility consulting is a good business to have these days, but I hope my work will be obsolete one day :).

I have been building software UI's for a decade and I absolutely agree.

I really don't get it. Basic web-pages are simpler, easier, faster, smaller, and provide data in a more manageable format.

Sure, there's some things that work better as an SPA, but those are extremely rare.

Basic web pages are less appealing to the mainstream, so they capture fewer eyeballs there.

Example: https://www.debian.org/ is a simple website, designed for broad accessibility, but it doesn't look 'slick'. If you were looking to choose an OS, this website wouldn't make you go 'wow'.

You are correct, but that's kind of my point.

To me, a website is a tool to get something done. I don't want a website that "wows" me. If you really want to wow me, then you are likely trying to sell me something that I probably don't want.

Accessibility is about letting people get shit done, not marketing to them.

If basic webpages 'sold', then marketers would use them. Webpages that 'wow' work, that's why people use them. It's not based on gut feelings either - there's lots of research in the area, for example A/B testing.

People say they don't want shitty news either, but the news that is most popular is pretty low quality, sensationalist tripe.

Your missing my point: If your goal for a webpage is to sell, then glwt. I get that they are a thing, but they are pretty useless to me, and good luck making marketing material that excludes people that want to give you money.

Webpages to me are tools, not marketing material. If you want to use them for marketing, then great! Just don't waste my time trying to turn a tool into marketing material.

Webpages that are meant to be useful (like Debian makes, etc) should be accessible, not promotional. The webpages I use are tools to get a job done, not to browse for advertisements.

> and good luck making marketing material that excludes people that want to give you money.

You're completely missing my point. Web pages are the way they are because they pull eyeballs. You are a single data point. You don't like slick webpages? Fine, you're an outlier. There are people out there that don't go online at all. They're also an outlier. You each find niches to cater to you, but you're not the majority case, and if you're interested in eyeballs for whatever reason, then dealing with the majority case is in your interests.

It's not just marketing either. Compare Arch's wiki (simple Mediawiki) to Debian's. Mediawiki is more complex than Debian's simple pages, yet it is far more readable. Even as 'just a tool, not marketing'[1], this heavier webpage is a better webpage.

[1] marketing isn't just selling stuff. for example, plenty of free software projects provide giveaways to conferences to raise their profile

Releasing content in to the commons for free //should be// an exception; however if someone else wants to add to that by providing captions in a standard format there should be a way of combining that to achieve a further enriched commons.

On the other hand, catering to ignored niches is a good business plan. If you can own a small segment, you can make more money than uselessly going after the center of where your entrenched competitors are.

Once the niche is conquered, this can be used to go after the less nichy segments.

See "The Innovator's Dilemma" by Christensen

The problem is that this niche is not stereotypically a very well paying niche. The intersection of high earners and disabled people is quite small (and the intersection between lowest 20% of income earners and disabled is quite large)n So, there's not really a lot of money (if any) to be made by focusing solely on this niche.

That's why the ADA and various government laws came into existence - nobody was catering to the niche since it wasn't lucrative enough.

I made good money in the compiler business by catering to niches that were completely ignored by the other vendors, large and small.

I'm also happy to get any accessibility problems in dlang.org and digitalmars.com fixed, if anyone wants to notify me of any problems.

That may be good for one individual business, but not good for the segment of the population that's disabled.

For example: restaurants. Take a world where restaurants have to comply with ADA regulations versus a world where it's expected that someone coming up with a business plan for a restaurant that caters to the disabled. It's the difference between a world where the disabled can go to restaurants or where they can got to a restaurant.

I live in a central European city full of restaurants that would be sued out of existence if they were in the states for ADA violations.

The ADA also means that a lot of hole in the wall establishments you see in Europe simply can't exist in the states. Society is poorer for it.

Is it? The disabled are part of society, too.

Where do you draw the line? Berkeley recently [0] had to pull tens of thousands of free education videos, because California law required that they be captioned, and they couldn't possibly caption tens of thousands of videos. So they just pulled them. I'm on board with an E for effort, but I find what I've described unacceptable.

[0] https://reason.com/blog/2017/03/07/berkeley-deletes-200000-f...

I think people underestimate the difficulty of getting this done in a University environment, but I also think people are too quick to side with Berkeley in this particular case. One of the points made by the DoJ was that Berkeley makes resources available to faculty and staff to do the captioning, so, in theory, the captioning could have been done as the videos were made (at least those which were made in the last couple of years), and it should have been done as they were posted (again, at least those posted in the last couple of years).

So, the problem is getting the faculty and staff at Berkeley (and nearly everywhere else) to take the requirement seriously. Many of the people working to make content accessible are not in a position of authority over the people creating the content. In some cases it's a battle to educate the people who are in authority regarding the legal requirements and the consequences of ignoring those requirements.

As much as I wish that content didn't have to be pulled because of accessibility issues, sometimes there is no other way forward.

It is worth noting that the Berkeley case was federal law. California law does make the requirement for state Universities to comply with the ADA more explicit, but the US DoJ isn't in the habit of investigating violations of state law. The DoJ's letter to the University cites the ADA and various portions of the US Code: https://news.berkeley.edu/wp-content/uploads/2016/09/2016-08...

> it should have been done as they were posted

This was not something they where required to do?The implication here is if you add requirements to free content, institutions just won't bother - the material was only released due to the low bar to do so in the first place.

> sometimes there is no other way forward

There is if you change the law, which is what I think parent was implying

This is where it can be difficult for people to understand the culture in which people are operating. From the outside, a public University is the ivory tower, but from the inside, it's just short of chaos. Sometimes faculty have direct access to post this stuff and either don't understand the requirements or don't care. Other times the staff who post this stuff don't have the support to stand up to faculty when they're told to post it.

The University is not removing all of their content, and they plan to continue to add new public content. It would definitely be a significant undertaking to make 20,000 audio and video files accessible. I'm sure the people responsible for the content will consider prioritizing that undertaking and possibly seek funding to get some of that content back out there.

Something else that is missing in their discussion of requiring University credentials to access the content is the fact that the same requirements apply to content they distribute to students, faculty, and staff. The difference is primarily in the likelihood they would be sued, and the fact that they would have a much easier time finding the funding to make something accessible if a student needed it.

Yes, you could change the law, which would theoretically benefit the people who could access this content, but would completely set back the few gains we've seen in web accessibility (assuming we're not also talking about everything else the ADA does).

Or you could set priorities for the content and fund the effort to make it accessible if it's really worth making it available. Fund research into improving automatic captioning/transcription, put your students on the path to better jobs, and improve the content for everyone.

That's hardly the biggest problem IMO.

What bewilders me most is the number of engineers and designers who can't empathize with how their users might interact with the unspeakably bad UX they're putting together.

Accessibility costs a lot of money and it's a perfectly valid observation that one can skimp on it for profit. You may call it immoral, I call it mathematics.

"Accessibility costs money" is a popular thing to say, but I'm not convinced that needs to be true. Okay, fine, captioning/transcribing videos I'll give you, but many of the other issues on this page can be addressed for a web page just by not breaking the defaults.

By default, links are underlined. For text content, many major browsers (Firefox and mobile Safari for sure) include a "reader mode" or similar where the user can adjust their own contrast - as long as you don't break your HTML and prevent the reader mode from working. Web pages are black and while by default, so anything with lots of bright colors is something you have to do intentionally. Animations are also something you have to add intentionally.

It really seems like just by not breaking the defaults (I.e., no work at all) your web page should do a decent (not perfect) job of being accessible. But we're so obsessed with making things shiny and swooshy that we've lost sight of that.

>"Accessibility costs money" is a popular thing to say, but I'm not convinced that needs to be true. Okay, fine, captioning/transcribing videos I'll give you,

Even that one is overstated. Sure, taking some existing video, and treating it as the only representation of the content, and then trying to transcribe it -- that's expensive.

But virtually every such video (esp. lectures) intended for conveying information is going to correspond to some text equivalent, and they can just post that in a place that's easy to find. Not only is it cheap, many users who can hear will prefer that version.

I think it's just a widespread attitude problem (for lack of a more refined diagnosis). It's not just big institutions, but discussion forums where some users will stubbornly link a video as their only citation, and won't even take the time to summarize the key point they think it's substantiating.

Can you elaborate why do you assume that such videos are likely to have a text equivalent?

You bring up lectures as example - I certainly don't have transcripts of lectures I've given, and neither do most professors/lecturers/whatever that I know; a few of them have most of their talk content in their slides but that's a horrible practice that distracts from effective lecture or presentation; in general all the non-audio material is intended to be a supplement to the talk and either are not meaningful without it (e.g. illustrations) or are not representative of the talk (e.g. a textbook chapter that's mostly about the same topic, but different). The recorded audio (if it'd be recorded) would literally be the only copy of the content of that lecture in existence.

The same applies for many other genres of video content. Most videos aren't scripted, and thus no script exists; heck, even videos that are clearly well prepared, staged and filmed with multiple takes (e.g. many of "youtube creators") generally don't have a script beforehand, the textual content is improvised on the spot and doesn't exist in written form unless it gets transcribed.

Interviews and podcasts are other good examples - as you say, many users who can hear would prefer to read it, I certainly would, but the obvious reality is that the textual version of those things does not exist unless it gets made, and making it takes much more time than recording the interview or podcast in the first place.

I was specifically referring to videos intended for conveying information, not entertainment. If professionally done, it has a script. If a college lecture, it has a corresponding textbook exposition. And you've seriously never written up any equivalent to any lecture content? It's all extemporaneous, all the time, and you can't spare the one time to write it down?

We don't have a tradition of teaching from a particular textbook (some universities do, some don't) so generally a lecture would include pointers towards chapters in multiple different textbooks, none of which match the contents of the lecture exactly (i.e. they're supplementary material, extra reading); and even if some lecture was based on a particular textbook, that text (unlike the lecture video/transcript) would be something that couldn't ever be published with a video because of copyright issues.

For my own teaching, no, I don't write up a "script" beforehand, I prepare content and plan for the topics and key points that I intend to cover in this lecture (and the next one, if we go quicker than expected), but the actual content heavily depends on how the students react, what they are (mis)understanding, what are the questions, etc. I don't consider well-scripted lectures a good thing, the lecturers that I know who are more into "reading off a script" give quite weak lectures IMHO.

Preparing in-depth textual material would seem a waste of effort as students would get the same content in the lecture. I wouldn't use it for other students since I'd anyway never give the same lecture twice - research is progressing fast, so my 2016 content is dated now and this year's content will definitely have to be rewritten (and the course restructured) next year. Admittedly, that's context-dependant - my situation is uncommon, there's a lot of teaching that involves lecturing the same introductory topic for many years off of a classic textbook to multiple larger audiences, and in that context your arguments would definitely apply.

But that's actually all offtopic.

The main issue is that none of this is equivalent to the transcript. Nothing of what you propose would make the video accessible to a deaf person. A textbook chapter or the lecture materials are alternatives to the video, it helps making do without the video, but doesn't make accessible the part of video where the lecturer is writing something interesting on the blackboard and explaining something that you can't hear.

I think you're the one who's demonstrating an atypical case here; most lectures aren't presenting on some rapidly changing topic, and even if they are, there is some Wikipedia text that is kept up to date.

>The main issue is that none of this is equivalent to the transcript. Nothing of what you propose would make the video accessible to a deaf person. A textbook chapter or the lecture materials are alternatives to the video, it helps making do without the video, but doesn't make accessible the part of video where the lecturer is writing something interesting on the blackboard and explaining something that you can't hear.

That's overstating it. The relevant metric is, if some deaf (or video-hating like me) person wants to learn the material, do they have a useful option? The fact that you can't make that specific video useful isn't relevant. Even pointing to some text that hits the same key points is major step up, and is trivial to do, and yet most refuse to do, just like most refuse to provide the one insight that they're making me watch a three minute video to hear.

By not breaking the defaults you make something ugly. By making something ugly people won't come again. Hence accessibility does cost money. Add the occasional ungrateful guy who will sue you because you forgot a caption and I get why people don't try to accomodate more to a niche that won't be lucrative anyway.

Black text on a white background with the occasional underline is ugly? God help you if you ever have to read a book.

By making things accessible to disabled people you also make them accessible to computers. This can be a great asset, for instance in automated testing (much easier with accessible ui) or indexing (with captioned video).

Making an accessible webpage to present some content is easy. Making an accessible web application is a completely different problem.

While on the surface this is true, the amount of work required as a developer to learn how to do this is insignificant compared to the total time required to really understand your craft. While you may have to spend more time testing accessibility if you want to use the latest features in the bleeding-edge versions of JavaScript and CSS (or whatever language you're using which compiles down to them) or WebAssembly, you'll probably spend more time on browser compatibility testing, too, with those technologies (or you'll take the IDGAF approach).

When you really dig into it, though, most of the accessibility issues for the UI of a web application are still among the things you have to know to present accessible content on a website. In either case, the important point is to know when you're pushing the boundaries so you can make an informed decision about your testing requirements. It's just another metric to consider, like browser compatibility or mobile presentation.

The cost of accessibility depends upon many factors.

If your design starts with accessibility in mind, the costs can be minimal. Keep in mind that the use of small and low contrast elements is a design decision. If the decision is made to use large and high contrast elements from the start, there should be no difference in the development costs. Making it possible to scale pages for pages where things are laid out spatially is far more difficult and incurs much more testing, yet costs can be reduced if the decision is made from the outset. Even the more difficult factors, such as captioning images and videos, can be eased if planning is done ahead.

The problem with arguing that it's 'mathematics' is that there are many variables involved in the equation. If the variables use values for poor planning and design decisions, the result is going to be a much higher cost. If the variables use values for solid planning and design decisions, the difference is going to be much lower. Of course it also depends upon the media being used, so it is very difficult to use that equation to make generalizations for all cases.

Just thought it would be worth pointing out some other costs we don't typically consider here besides development costs is litigation costs...

The most troubling wave of ADA litigation is just beginning to crest. The ADA’s next frontier—and next target of litigation—is the most innovative segment of the domestic economy: e-commerce. The latest trend in ADA litigation is suing commercial websites that aren’t sufficiently accessible to the disabled—because, for instance, they lack assistive technologies for the blind or hearing-impaired. Even though the ADA was enacted before websites became ubiquitous, many courts have interpreted the term “public accommodation” in Title III to encompass the Internet, and will entertain suits applying to it. In 2008, Target paid $6 million to settle a class-action suit brought by the National Federation of the Blind charging that the retailer’s website was allegedly insufficiently accessible to blind customers. Target also paid nearly $4 million in costs and attorneys’ fees to the plaintiffs. Why would Target pay $10 million to settle a case when the law is unsettled and its culpability far from certain? Class-action lawsuits are particularly expensive to defend, and because the potential liability is aggregated to reflect the entire class (all blind consumers, in this example), the financial exposure for a firm if it loses in court presents an unacceptable risk to most corporate defendants. https://www.city-journal.org/html/ada-litigation-monster-151...

This also happened with Netflix over captioning: https://www.nad.org/2011/06/16/nad-files-disability-civil-ri...

Netflix now does the right thing and captions everything.

Was there a tradeoff involved there? Is there content that is not in the Netflix catalog today because of the costs of closed-captioning the rest of the catalog?

If that is the case, are Netflix's customers better or worse off, individually and collectively? Is society better or worse off? (I don't have a clear answer, and I'm not sure there's an objectively clear answer, but I do think that there are tradeoffs in these discussions that are sometimes hidden/ignored.)

I am not aware of any content that got ditched because of this. I imagine only a Netflix admin type person from back then would know something like this.

They only have to provide it, if it originally had captions. They didn't have to caption stuff that never had captions to begin with, but they went ahead and captioned everything, pretty much, and if you look at their new self-created shows, they are typically multi-lingual(multiple voiced languages) and multi-captioned (multiple written languages).

I would say, it forced them to deal with becoming a global provider and not US/English centric like they were, which I think we can say has been good to their business overall :)

Except that makes things easier for everyone else too. For example I often can't watch videos at work. If that's the only way to get information about your product or fix to an issues you might as well be speaking Martian in that video I can't use it.

The engines behind Alexa and Google Assistant have decent enough accuracy, I'm not sure why there couldn't be a product to run the audio through those to get subtitles.

When there's not a single audio source for the main speaker, that becomes a problem, but that's the fault of the people recording it for not getting a clean audio source. However, in practice, this is rarely going to be the case for something like a product video because the studio will have recorded it correctly in isolated sound rooms.

We're automatically reverting to "a human should do it" with adding aria tags, video subtitles, and screen readers. That is the problem because developers often have many other things to worry about -- database, security, scalability, maintenance, documentation. Accessibility gets put right beside those and it's no wonder developers don't prioritize it.

It doesn't work well enough. We're no where NEAR it working well enough. People invent new words (product names, technologies), the text isn't synced to the video, there is background noise, the speech-to-text engine doesn't have the context of the video to guess which homophone is correct.

Technology is not the magic bullet to fix everything. We're no where near that place.

In the mean time, I still can't use your site. And neither can people with any number of other issues (medical, environmental, simple bandwidth, etc).

Don't pawn it off on tech, people need to take responsibility for this kind of stuff.

More importantly, accessibility is hard and complicated.

There aren't good tools for collaboratively captioning videos ala a Google Doc. ARIA is a massively complicated spec that is much harder than <font color="red"> tags.

There are simple and easy tools for making inaccessible content (Flash was a big one). It is much, much harder to make accessible content, and it shouldn't be that way.

Perfect compliance with Sccessibility can be hard and complicated; but have another look at the original article. The things being asked for are not hard or complicated.

Making no effort just because a perfect effort would be hard comes across as a cop out. Would we accept this so easily if we were discussing testing?

If you can push back on $BOSS, certainly. But, frankly, it's way too easy to make inaccessible content.

It should be impossible to make inaccessible content. We could have browsers and/or compilers that just break the whole page if it isn't accessible.

I dunno, satisfying both "not low contrast" and "not bright colour schemes" together, along with a good design in general, requires a moderate amount of knowledge.

I'm not arguing that it shouldn't be done, but the idea that it's effortless is simply not true.

Financial perspective is an issue but this also.

Its complicated and its not directly related to your core work/competency. One uses a lot of frameworks/tools to build on top of what if they are accessible and you can't build the thing completely on your own? Its quite to get started to be sure which one is just some regulatory checklist and which ones actually matter. I have trying to be knowledgeable on this but compared to normal web development informations are both sparse and harder to grok. And its also much harder to test/validate than just checking if my script/css is working/looking right. I have done very little screenreader tests but it somethings they are terribly good at working through and some things break too easily. It might make perfect sense for an accessibility expert but can be very non-intuitive for a regular developer.

If this is truly the case then website accessibility will end up going the way of physical disability accommodations -- legally required.

It has in some cases. In the US, The DOT requires it for airline websites.

Ultimately you're mostly talking about the cost of digital typography.

...screen readers just need to get better.

TFA doesn't relate any tweets about screen readers, and identifies many easier tasks as being more important than screen reader compliance.

It would be nice if web authors had a tool for that compliance that wasn't expensive and obscure like the readers themselves. (If they weren't obscure, we'd just file some PRs and be done with this.)

NVDA [1] and VoiceOver (iOS and MacOS) are free.

[1]: https://www.nvaccess.org

In my experience companies don't want to pay for it up front. So they end up paying much more later (after the lawsuit).

How many times have you selflessly made your product accessible? It's easy to point fingers and judge people, but it is genuinely difficult to serve people with every conceivable disability.

Do you optimize the prose on your websites for people with IQs under 75? How exhaustive is exhaustive enough?

I've spent considerable effort on making things accessible in my life, and I'm sure it can cost you your entire product when you're just starting out.

Edit: I think the limitations apply to products which are marketed to the general public. If you know your existing users are struggling with an accessibility bug, then you know what you have to do.

Writing documentation for the equivalent of 8th graders is pretty common for any writing, so I made a habit of doing that a long time ago. It's like coding - if you're too clever, it's hard to understand and read later.

How hard is it to really use reasonable font sizes and contrast ratios? How hard is it to add a few more tags to your site? How hard is it to caption videos as you're making them (most time you're working from a script anyways).

If you're not starting these good practices from the beginning, yes, it will be harder to do all at once later. But if you start doing it from the beginning, it is pretty easy to do. Just like testing. Just like coding standards. Just like design patterns.

It's easy to accuse people of pointing fingers and judging people, too, and genuinely difficult to make high-quality, thought-provoking Hacker News postings.

(Your comment would have been much better had you left out the rude accusations.)

Reply on news.ycombinator.com