Poll: Do you use access keys?
I felt it was about time to bring up the topic in one of my oldest drafts (written in May 2005) which, for some reason, never saw the light of day.
What are access keys?
Basically, access keys is about connecting keyboard shortcuts to certain HTML elements, such as links, search form fields in a web page etc. Paired with a certain trigger key, different depending per operating system (normally Alt in Windows, Ctrl in Mac), it can be used in any web page. It can look like this:
<a href="#main-content" accesskey="s"> Skip to main content</a>
The reason they exist is to offer another navigation option for people with different disabilities, such as screen reader users, people who can only use the keyboard but not a mouse, developers like me who are obsessed with keyboard control. There are some fairly standardized access keys, such as:
- S: Skip navigation
- 1: Go to home page
- 4: Search (set focus to search field)
A more detailed list with different alternatives can be found in: Accesskey standards, and various government accessibility guidelines set their own standards.
As you probably have figured out by now, access keys are bound to collide with other operating system-, web browser- or web browser-extension shortcuts. While attempts have been there to offer the user to define access keys (Greasemonkey version), personally I think the same access keys should go for all web sites, or the cause is somewhat lost.
And that kind of leads to the gist of the problem: most web developers don’t know about access keys and/or don’t think it’s worth the hassle of offering them to the end user, due to above mentioned collisions and inconsistencies. Me, I use the most common ones like skip, search and home page, but otherwise I tend not to include them (unless it’s for some sort of government web site with certain regulations).
A little poll
Out of curiosity, I thought I’d introduce my first poll-in-a-blog-post to see whether you use access keys in your code (recommended: visit my web site to vote). Comments about the existence and necessity of access keys are of course also very welcome.
I've considered it a few times, but never actually implemented it.
I'm quite happy with the Ctrl-E to edit that exist in Trac's and MediaWiki's wiki software. But I never put that kind of code into my own apps.
At our recent Refresh meeting we had a blind person come in and share with us their experience on the web. It was a great meeting where I learned a lot, including most vision impaired users who have screen readers feel that access keys get in there way and they just ignore them. I guess there are other type of shortcuts or something built in the readers. I may be wrong I can't remember to much about that but I know she said that other disabilities may use them more.
No, for the exact reasons you outlined. I absolutely hate it when they collide and I feel the lack of standardization as well as discovery from within the browser makes it incredibly hard for the above-average to expert user to fully utilize them. Requiring them to be used in combination with modifier keys makes it a lot more complicated.
I make one exception and that's when I'm designing an interface that's more of an application and needs to consider quick and high frequent use. In such cases I tend to go the Google Reader route and assigning shortcut keys to single buttons. They only get in the way when using "Find as You Type" in Firefox (which I do, unfortunately ;-).
I did use accesskeys for some years but stopped a couple of years ago due to lack of standardization. I liked the idea that the accesskey is the first letter of the link name – thus the keys could be used without learning a system beforehand…
For public facing websites, I always imagine them getting in the way of visitors using the website, since there are so many collisions that developers have to get very creative to get enough of them to work. This puts the audience at great disadvantage as they have to learn a lot of weird shortcuts for a site they maybe only visit once.
For public facing applicaitons and internal applications, it is certainly interesting, unfortunately I haven't been able to include accesskeys since 2004 🙁
I voted "never".
I've known about access keys for several years. Used them in the past on websites, and abandoned them a couple years ago when they collided with other shortcuts.
Hi Dan! Dan and I were at the Refresh Detroit meeting last week where Cathy McAdam, an advocate for assistive technology and a screen reader user, spoke about access keys and other methods for improving the experience for all website users.
It would be nice to some standardizations emerging for access key usage.
Visual impaired people then, someone may ask? Well, that's another story…..
I'd love to say "never" but I have had clients who have insisted upon them, despite being shown the arguments against them.
In particular, I believe they are required as part of the Shaw Trust certification. I'm meeting them in a week so hopefully I'll be able to have a chat with them and find out their reasoning.
I never used access keys before cause I am thinking "It will not work"
Alt + key can trigger many actions in browsers. I believe we should push till the standardization issue is solved!
We let our users define accesskeys on our forum for various functions (like submitting forms; I think [key]+S is mostly used for that) or links (such as their posthistory). I myself use my own defined accesskeys all the time (but then I also spent a lot of time on our forum).
Recently the usefulness of accesskeys regressed; first when Firefox changed the default accelerator key, and with version 3 accesskeys defined on non-visble elements stopped working (see bug 419059).
IE has always been a pain wrt accesskeys defined on links; when triggered it just focusses the link but does not follow it.
It has been suggested in before mentioned bugzilla bugreport that perhaps HTML5 should define accesskey behaviour; let’s hope that it will or come up with some better alternative, because in essence I think it is a very good feature (maybe even more so for mobile) but it is more and more becoming absolutely useless because of browser implementations, differences between them, and lack of (easy) configurability by the end-user.
Whenever I think about accesskeys, I'm reminded of the usability rule that â€œPeople always spend more time on other sites.â€ So sans standardisation, there are very few sites where itâ€™s worth anyoneâ€™s while to learn and use accesskeys…
Apart from that, picking appropriate keystrokes (which work cross-platform without collisions) is a real pain.
Perhaps we need a psuedo-standardisation body along the lines of the Microformats group, to work out a standard set of keystrokes.
I tried it once, but was quick to abandonne the use of acceskeys when the client rang to inform me the site didn't work properly because his shortcuts didn't work. Me explaining about collisions didn't really help so I quickly "fixed" the problem and left them alone from then on.
With proper standardisation I could be convinced though…
I think it was a great idea that just didn't stand the test of implementation – there's no universally agreed set of shortcuts that are "safe" and won't clash with the OS or browser. If those conditions could be met then yeah I'd probably put them in.
Working on governmental sites in Sweden pretty much demands accesskeys to be defined.
In other projects I've not used them. Thinking of it I realised I haven't spent any thoughts about them since I don't "see them". Anyways – I found this Custom-CSS block for FF to make them visible: http://blog.andrewbeacock.com/2007/06/firefox-hac…
What would the safe set of Accesskeys be if I did change my mind and consider using them?
My understanding is that between OS and browsers only about 2 keys aren't used by someone for something. It's probably a bit arrogant to hijack their usual behaviours and implant our own, so regardless of even accessibility issues it's just one of those good ideas that never worked IMO.
Secondly, where is the standard for creating accesskey sets? So users have to relearn a new set for every individual site they visit. That, in the end, doesn't really work. Even if we give them personalised instructions… on the flip side, if we let them allow accesskeys then do they honestly know or care to learn how to implement and maintain that system anyway?
So I'm not sold on accesskeys. Good idea badly implemented… just my 2 cents… gotta run to the theatre… ciao.
Great discussion, people! Thanks for some fantastic input!
Probably, it’s too late for standardization, I think. It’s a great idea, just access keys themselves, but ion practice, it has gone too far. And changing anything by now would probably upset more people, than helping the ones who need that.
Only existing key control I see in web pages is like Google do it (with at least the same keys across their applications) and like Tino mentioned where they are custom made for a specific web site.
Very interesting about someone being blind and how they see it! It would be great to collect a number of people with various disabilities and try to understand each person’s point of view on this.
Thanks, nice tip!
No, I do not use accesskeys when it comes to public sites which should be as accessible and usable as possible. I get answered "sometimes" because I've done some development on intranet webapps, where the target audience is a small group of people – usually less than 50 – and there could be some education and workshops about the interface. In that forum, access keys is safe enugh to use as a tool to be more productive.
But when it comes to sites online, available to the big mob – no. Just no.
I share Anders' view on this. On the web, I never ever use them for reasons mentioned by Steven Clark. But on intranets it's another matter. Since I tend to deploy more application-like content on intranets, they need to mimic their cousin counterparts. They need accesskeys to do that properly. Users demand it every time 🙂
I use them. Currently using http://2bweb.biz/accesskey/accesskey.html
I am somewhat visually impaired. Far from beeing blind, just significant problems, perhaps due to age (51). I found that some access keys could help. The problem is the missing standard. But the idea of at least a standard access key to the main content is a good idea. I think i’ll use that. Thanks for the idea.
The problem with web design is that most (nearly all) web designers are graphics designers. And from that point of view access keys are irrelevant.
BTW: I’m using NoScript. And sometimes i switch off styles and images to get at least some content, which else would be lost in blinking and flickering ads, bombasic decoration and an unreadable font size (at least for me). Switching it all off sometimes makes a web page accessible for me.
But of course i’m not the average surfer. I’m a software developer since ages, and so quite IT affine, and i know my browser. Most surfers do not know their browser. Most (nearly all) surfers depend on the default settings of their browser and the nonsense many web designers produce. Even people with disabilities much more serious than mine do not know their browser. I know of a man needing a font size of 72px and a magnifier to be able to read something. And he did not know that the browser offers settings for the default font size.
So back to access keys: As long as they are not really wide spread and a real standard exists, very very people will know about them and know how to make use of them. Even people with disabilities won’t use them because they simply do not know how to, and because there is no reliable standard.
[…] Robert Nyman gibt’s gerade eine Diskussion Ã¼ber “Access Keys”. Und da hat er mich auf eine Idee gebracht. Ich […]
Thanks for sharing!
Very interesting to hear!
I also believe that most people with different disabilities don't know too much about technology, making it an even bigger problem to accomodate to them.
All we can do now, I think, is teach them about the options and spread the word.
Well, probably very few of those people who would need reading this will in fact read this. Web pages about web design are for IT affine people, not for the average surfer. So f.ex. this artikle will be read by people aware of the problems and knowing their browser. But it should instead be read by people in need to exploit the features of their browser.
Perhaps another approach could help. The browser developers could add a guided intstallation tour with special focus on accessibility features and help texts describing for what purpose the feature is, how it might help, and how to make best use of it. This guided tour should always automatically start when first time using the browser or first time using a new browser version (not a minor version upgrade, but a makor version upgrade). This way everyone would at least get an idea of what a browser does.
I really like that idea! You should talk to web browser vendors and suggest this – I'm sure they appreciate the input!
o.k., done (at least for firefox): http://forums.mozillazine.org/viewtopic.php?f=7&a…
[…] Poll: Do you use access keys? […]
I'm surprised to see such a high percentage of accesskey users. I thought for sure it'd be zero.
I use them on my site but use a version of Dan Champion's user-defined accesskeys script modified for WordPress to let site users set their own accesskeys.
I used to define accesskeys, when I was naive enough to think they were truly useful. Now I think they're a useful idea, but so badly implemented as to make them useless in practical terms. This pretty much says it all: Using Accesskeys – Is it worth it?
We used to use them a few years back, but after reading all the research and first hand experience of issues we decided not to use them on any of our sites of ones we design.
We also wouldn't recommend them to our clients.
We also advise of the issues they can cause to any of our clients sites we test/audit.
Another much needed post.
I personally donâ€™t use them, though I really would like to. We do need a standard, at least a de facto one. Maybe this conversation would go further then a simple comment thread.
Wouldn’t it be interesting to find out how many users *used* accesskeys, rather than querying how many developers implementing them?
(although that’s really interesting as well!)
Thanks for your comments!
Good input, that question is also very interesting!
You have raised very relevant issue.
[…] Do you use access keys? gyorsbillenty?k a HTML kÃ³dban […]