RE: SC 2.4.5, meaningful link text

The editor's note on 2.4.5 asks us to comment about how closely link text
should be associated with text describing the destination. Below I will
provide the requested comments and then a recommendation:

 

In order for a person using JAWS to access the words around a link they need
to TAB to the link and press INSERT + LEFT ARROW or INSERT + RIGHT ARROW. In
Home Page Reader (HPR) it requires SHIFT+CONTROL+LEFT (or RIGHT) ARROW. In
WindowEyes it requires the same SHIFT+CONTROL+LEFT (or RIGHT) ARROW.

 

My experience as someone who works in the field of ergonomics, as well as
providing Assistive Devices accommodation, is that the incidence of
repetitive strain injury (RSI) is higher among screen reader users than in
the general population of computer users. Forcing blind users to make
unnecessary extra combinations of keystrokes, is not fair I would say. This
may not be as big an issue for our young blind users. But over the years the
extra wear and tear on their elbows, wrists and hands will often take their
toll. I have seen this. Assistive Technology is supposed to overcome
disability, not cause it.

 

It also takes extra time to TAB and twist the hands to hit INSERT + LEFT
(RIGHT) ARROW or CTL +SHIFT +ARROW.

 

There is nothing more important on the internet than links, it is what makes
it the internet. Sighted users effortlessly visually skim the page  for the
links they want. (often with no key strokes)Why would we want to force blind
users to risk cumulative Repetitive Strain Injury over the years (hundreds
of thousands of extra keystokes) when it is such a fundamental part of what
the web is all about? I don't think this is about taking the links out of
context, anymore than zooming into a small part of a page with a screen
magnifier is taking information out of context. I don't know why we would
want to make it harder for AT users to compensate for what sighted people do
effortlessly.

 

It is not hard to create meaningful text links, as long as we allow
exceptions for arrays of links. Explicit meaningful text is priority 2 in
the 1.0 Guidelines. I think there will be quite a big price if we let this
slide off the table in the 2.0 Guidelines.

 

Therefore, for these reasons and others, I think we should be fostering a
culture of meaningful link text. 

 

*Recommendation* 

 

SC 2.4.5

 

<current>

Each programmatic reference to another delivery unit or to another location
in the same delivery unit, is associated with text describing the
destination. 

</current>

 

<proposed>

Each programmatic reference to another delivery unit or to another location
in the same delivery unit, is programmatically associated with text
describing the destination, unless it is part of an array of programmatic
references to different versions (or views) of the same information. 

</proposed>

 

Regards

David MacDonald

 

 

.Access empowers people

            .barriers disable them.

 

 www.eramp.com

 

 

-----Original Message-----
From: w3c-wai-gl-request@w3.org [mailto:w3c-wai-gl-request@w3.org] On Behalf
Of Ben Caldwell
Sent: Wednesday, January 04, 2006 5:39 PM
To: Jason White
Cc: w3c-wai-gl@w3.org
Subject: Re: SC 2.4.5, meaningful link text

 

 

Jason White wrote:

> On Mon, Jan 02, 2006 at 12:45:41PM -0600, Gregg Vanderheiden wrote:

>>Also - the question is how helpful.  And why couldn't AT be programmed to

>>allow users to get information around a link with a simple keystroke for

>>those cases where the link all by itself did not give them enough

>>information.

> 

> Do GUI screen readers make this hard? Under all of the Unix text browsers
I

> use, and with both braille and speech assistive technologies it's trivial
to

> read the line containing the link, the lines before and after, etc. As a

> result, this has never struck me as a concern. At most it's a couple of

> seconds of extra work to read the context. I tend to read the text of

> unfamiliar pages anyway, rather than just reading links, so for my usage

> pattern the problem rarely arises. With familiar pages I use the "text
search"

> function of whichever user agent I'm running to get straight to the
desired

> point without having to navigate to it.

> 

> I suspect it's the kind of problem that affects some user agent/assistive

> technology combinations more than others, and some people more than
others.

 

My understanding is that reading the text that surrounds a link in GUI 

screen readers is not at all difficult. It is true that when a user 

pulls up a list of links on a page, surrounding text would no longer be 

available, but I don't believe this is something we should be concerned 

with as this is a case of a UA feature that re-displays the content 

outside of its original context.

 

David, can you clarify where you recommend including this? I assume your 

proposed text (Provide meaningful link text, unless the link is part of 

an array of links to different versions (or views) of the same 

information.) is a technique, but it is phrased like a success 

criterion. If the former, would this be advisory or sufficient?

 

-Ben