Re: SVG 1.2 Comment: Accessibility: event handling

Chris,

unfortunately I've not expressed myself sufficiently well, which is an 
habitual problem.

Whilst the specification fails to take account of the needs of keyboard 
users, Authors create user agents that fail to be usable via the 
keyboard, and yet claim they are for instance SVG1.1 compliant.

We are then in the unfortunate position that not only are these UA not 
accessible, but furthermore there is no usability or accessibility 
information generated for the next specification.

Please bear in mind that it was 'feedback' that led to the the change 
in event handling you refer to.
The international community, would be expected to have a stronger voice 
than the disabled community.

You failed to address my query as to whether there is a  UA that is 
keyboard accessible, though I would expect there are some.

regarding the UAAG, surely if keyboards events  are hopelessly broken, 
it might be helpful if references to them were updated? Or if the link 
from the spec to the UAAG was removed, if you dont wish UA to support 
current accessibility guidelines.

regards

Jonathan Chetwynd
http://www.peepo.co.uk     "It's easy to use"
irc://freenode/accessibility
On 19 Nov 2004, at 21:17, Chris Lilley wrote:


On Friday, November 19, 2004, 9:15:05 PM, Jonathan wrote:

JC> Chris,

JC> I'm more than a little concerned regarding the accessibility of 
current
JC> SVG user agents, and question why this lack of progress is not
JC> informing the development of the current 1.2 specification.

I'm primarily concerned with the specification, of course; and ensuring
this gives adequate direction to implementors of user agents and AT
add-ons.

JC> You state: "any pointing device which can give a position will 
work."
JC> this assertion is easy to make, but perhaps harder to obtain.

Well, its important to make the assertion rather that 'the events are
for mice only' for example.

JC> In a recent exchange on irc:mozilla/svg it was suggested that 
"svg1.1
JC> does not specify key events", which in it's way is true.

Its certainly true. Its not just 'suggested'.

JC> (the implication being that therefore it is not a requirement.)
JC> UAAG states:  " Ensure that the user can operate, through keyboard
JC> input alone, any user agent functionality available through the user
JC> interface" .

Again, you are mixing up specifications and user agents.

JC> This may partly be explained by the rather ineffective: "an SVG user
JC> agent should conform to UAAG." - should - may have been intended as
JC> sufficient, but in this instance the flesh is weak, as we are now 
some
JC> years late.
JC> asv is no better, as the transition from browser to viewer is 
broken,
JC> and the number of keyboard navigable pages ~=0.
JC>   - It maybe that there are other UA that have keyboard access, 
please
JC> advise me. -

As you correctly observe, SVG 1.0/1.1 has *no* keyboard events, because
the DOM2 keyboard events were removed from he DOm specification due to
being hopelessly broken from an internationalization point of view and
thus inaccessible to most of the world population.

SVG 1.2 uses DOM3 events, which

a) do give keyboard  and text access
b) are well internationalised
c) abstract out the inpt device from the text it produces and this,
allow alternative accessible input devices to easily fit in with
unchanged content.

Perhaps you could tell *us* about some accessibility aids that function
via the DOM, or even get some manufacturers of such devices interested
in

JC> The significant change in event handling is not reflected in the
JC> current UAAG or indeed the overdue SVG Accessibility AT and UA
JC> guidelines. Does the 1.2 specification indicate the significance of
JC> accessibility in an appropriate manner, given the response to the 
1.1
JC> guidelines, bearing in mind the lack of progress on updating these
JC> accessibility guidelines.

Not sure what you are trying to say there.

JC>  It isn't possible to refer to them, because
JC> they don't exist, and furthermore there haven't been usability 
studies
JC> to inform their development.

I'm fairly sure that devices such as head-mounted pointers, eye
trackers, etc exist having seen videos of them in use. If they can give
a screen position, thats all that is needed.


-- 
  Chris Lilley                    mailto:chris@w3.org
  Chair, W3C SVG Working Group
  Member, W3C Technical Architecture Group

Received on Saturday, 20 November 2004 08:35:43 UTC