Several Reasons to Avoid Accesskeys
For the most part I’d expect this post to be preaching to the converted, but I’ll say it anyway because I see the occasional implementation of accesskeys out there warranting a periodic caution to think twice before using them. I think the major problem is that older posts such as the 2003 A List Apart article Accesskeys: Unlocking Hidden Navigation by Stuart Robertson are still available on well respected standardista websites. The fact is that accesskeys have always faced several issues including:
- they are defined at the site level so users need to discover new accesskeys on each website
- according to WATS a few years ago all but 3 keys are already claimed and even those keys may still cause conflicts
- the last point also works under the assumption the user has a QUERTY keyboard
- attempts to standardise shortcut keys haven’t provided the promised consistency of shortcut keys across sites
- accesskeys are not in themselves visible to the user on reaching a website
Overall, given issues with accesskey and that they can pose more barriers than they pull down I’d strongly suggest avoiding them entirely. Its a shame because its a good idea implemented like a sock full of sand. I at least would like anyone who is thinking of implementing them to consider the ramifications. The last comment on Dave Shea’s 2003 post is also interesting - internationalisation. Done (shortcut D) turns to Fin in French. So how would D make any sense? Good point.
The user defined accesskey PHP hack looks kind of interesting but note Robert Nyman’s initial comment - how many people even know about accesskeys let alone feel like defining them? Seriously?
It would be nice to see A List Apart write some balanced commentary at the end of that article from 2003 stating buyer beware on this subject. Even if every browser implemented this consistently and correctly we would still have the conflicts with existing keys and inter-site consistency of implementation. No, accesskeys aren’t quite the bee’s knees after all.
