Re: Accesskey - spec proposal

Maciej Stachowiak <>, 2007-07-03 10:22 -0700:

> On Jul 3, 2007, at 4:08 AM, Charles McCathieNevile wrote:
>> On Tue, 03 Jul 2007 12:30:20 +0200, Maciej Stachowiak <> 
>> wrote:
>>> It also seems that accesskey would not work very well on mobile devices, 
>>> which often have limited keyboards.
>> The WAP forum (now OMA) obviously disagreed with you, since on WAP 
>> browsers accesskey (although restricted to numbers 1-9) is widely 
>> interoperable.
> 1) That's obviously not interoperable with the use of accesskey in general 
> web content (as opposed to waled garden "mobile web" content), since the 
> few real sites that use accesskey do not restrict themselves to 1-9. 

That depends on what you classify as real sites. The numbers of
HTML pages that use accesskey markup is conservatively on the
order of at least the tens of thousands. The number of users
accessing and those sites daily is certainly on the order of

> 2) How does this work on phones that have a full alphabetic keyboard, but 
> where you have to hold down a special numeric modifier to type numbers in 
> any context but the dialer? (For instance, my Nokia e61). And does an 
> unmodified 1-9 still act as a rapid access key when typing into a text 
> field on such a phone?

Honestly, I don't know. Such devices are not in wide use in Japan,
maybe for the reason that they are at a usability disadvantage as
far as being practical for use cases in Japan as far a mobile
usage goes -- because they require the use of two hands.

One of the main and most important environments that devices for
the Japanese market need to address is the case of commuters using
devices on crowded trains on their long commutes to and from their
workplace. A 90-minute train ride on one-way to work each day,
standing for most or all of the way, on a crowded train, is not
considered abnormal at all, and 2-hour+ commutes are not uncommon.

Standing commuters only have one hand free (the other being used
to hold on to a strap or other support), so they are restricted to
making use of what they can hold and use with one hand -- be it a
small book (even first-edition novels and bestsellers in Japan are
typically published in paperback form, and on top of that, in a
small format that is designed for "one handed" reading) or handset.

> 3) How does work on phones that are all touchscreen and don't even show an 
> onscreen keyboard most of the time?

The primary real-world use case for accesskey is for handsets that
lack a pointing device and are normally expected to be operated
with one hand. On such devices, the only directional navigation
control control is a so-called 5-way, which is essentially the
equivalent of the set of arrow keys on a keyboard.

Scrolling and trying to activate links using a 5-way is an
expensive operation for users, and something that good site design
should prevent them for needing to do.

So sites that intend to be usable and navigable with one hand use
access keys as a way of mitigating the user expense and
frustration with navigation using only a 5-way.

A touchscreen, on the other hand, has a pointing device -- your
finger -- so it obviates the usefulness of access keys; you can
just touch links directly to activate them, so there is no
navigation expense to you (though there is the issue that you're
not going to be able to use that device in a typical crowded train
environment like the one I described above).

> High-end phones are usually in categories 2 and 3, so it sounds like the 
> types of phones capable of rendering real web content well haven't been 
> considered in this design.

High-end handsets in Japan do not fit into your categories 2 or 3.
In fact, the average typical handset currently being sold in Japan
is probably a significantly higher-end in terms of device
resources and features that most so-called smartphones in North
America. Most handsets has QVGA-size screens (240 by 320) -- and
handsets with WVGA (480 by 800) screens are available, with every
reason to suspect they are going to become the new standard size
here eventually. The handsets all have built-in GPS, faster CPUs,
more RAM, gigabytes of removable storage (using SD and MiniSD
cards), sophisticated sound hardware and software capabilities for
storing and playing music files, built-in cameras of probably a
minimum of 2 megapixel (and handsets with 3-, 4- and 5-megapixel
cameras are available.

All that said, they still just have keypads and 5-ways as their
user input interfaces, and that is not going to be changing any
time soon -- because that interface is the only kind that
currently meets the requirements of the usable-with-one-hand use

> WML is not the web. Has anyone done this on a mobile device for HTML?



Michael(tm) Smith

Received on Wednesday, 4 July 2007 16:35:48 UTC