W3C home > Mailing lists > Public > public-html@w3.org > April 2011

Re: longdesc - beside the box

From: Leif Halvard Silli <xn--mlform-iua@xn--mlform-iua.no>
Date: Tue, 26 Apr 2011 16:32:45 +0200
To: Benjamin Hawkes-Lewis <bhawkeslewis@googlemail.com>
Cc: John Foliot <jfoliot@stanford.edu>, HTMLWG WG <public-html@w3.org>
Message-ID: <20110426163245797253.31c7f9db@xn--mlform-iua.no>
Benjamin Hawkes-Lewis, Tue, 26 Apr 2011 13:55:58 +0100:
> On Tue, Apr 26, 2011 at 12:22 PM, Leif Halvard Silli  wrote:
>> Benjamin Hawkes-Lewis, Tue, 26 Apr 2011 10:16:32 +0100:
>>> On Tue, Apr 26, 2011 at 2:57 AM, Leif Halvard Silli wrote:
>>>> iCab does show a "default visual encumbrance" for images with
>>>> @longdesc.
>>> The user has to take a special action (hovering over the image) to
>>> display the encumbrance (a cursor change), so it's not "default".
>> That's the same for many links: until you hover above them, you don't
>> see it is a link.
> @longdesc is hidden metadata by design;

The 'Techniques for User Agent Accessibility Guidelines 1.0' from 2002 
did not design it as hidden:


> hyperlinks and most aspects of
> microformats are visible data by design.

While, as you said, it is considered good to use underlining for links, 
it is probably more important what happens with the cursor. Not least 
because links where the image is the sole content do not get any 
underlining by default.

Thus an image that is a link is in fact (easy) discoverable meta data 
and not visible data - by today's design.

(The blue frame around images disappeared because it was considered 
ugly, but it can in fact still be seen on img elements with @usemap in 
Internet Explorer [in some modes, at least].)

> That you can write stylesheets
> that obfuscate visible data does not undermine the principle that showing
> data should be preferred as it makes error discovery more likely.
> Instead, it merely shows that the hidden metadata antipattern can be
> implemented at the level of skinning as well as semantic markup.

If it wasn't designed to be as hidden as you claim, then the argument 
doesn't stand. But it sounds as if we agree that there should be a way 
to bring the user's attention to the fact that there is longdesc link. 
We probably also agree that HTML4 was underspecified.
>>> If we accepted iCab's behavior as a "default visual encumbrance",
>>> we'd need to reject all Laura's examples of long descriptions with
>>> "No Forced Visual Encumbrance or Default Visual Indicator".
>>> http://www.d.umn.edu/~lcarlson/research/ld.html#noclutter
>> Sorry, but I don't quite follow. I'd say iCab's behaviour is
>> compatible with what she writes.
> Interpreting iCab's behavior as a default visual encumbrance is not
> compatible, because then her examples of webpages without a default
> visual encumbrance would in fact be examples of webpages *with* a default
> visual encumbrance (in iCab).

OK. But it is up to Laura to tell whether she thinks a cursor behaviour 
still is acceptable. She also, btw, points to the 'Techniques for User 
Agent Accessibility Guidelines 1.0' on here CP. And I know she has 
recently tested iCab, so she can speak for herself.

In principle I agree with Steve: "Showing a indicator next to or inline 
with the image when the image either recieves focus or moused over is 
not a 'forced visual encumberence'." Especially since he says it can be 
regulated via CSS. I believe Laura too also agree that it should be 
possible to style via CSS. Wether there by *default* should be visible 
indicator on the picture itself - that's a detail that we still need to 
discuss more. I myself am uncertain. I think it is enough if if it easy 
to make the indicator displayed.

> As I don't interpret iCab's behavior as a default visual encumbrance, I
> find it compatible with Laura's examples.
>> I'll also say that, when looking at the longdesc bug(s) in Mozilla, it
>> seemed exactly the 'context-menu' cursor issue caused a lot of fuzz:
>> https://bugzilla.mozilla.org/show_bug.cgi?id=258960
>> http://html4all.org/pipermail/list_html4all.org/2011-April/001109.html
>> (FWIW: may be it is best to not use the 'context-menu' cursor but
>> rather, as iCab does, use a similar 'there is an linked document for
>> this item' cursor..)
> I don't think the precise hover styling makes much difference to whether
> @longdesc is visible data or hidden metadata.

On that we disagree then.
leif halvard silli
Received on Tuesday, 26 April 2011 14:33:15 UTC

This archive was generated by hypermail 2.4.0 : Saturday, 9 October 2021 18:45:36 UTC