Re: accessibility review and 4.0 release

Jason wrote:

> 1. Since my original message on the topic, I have revised my braille media
> type proposal. As stated in my Braille CSS requirements document, I have
> suggested that the following two media types be defined, parallel to the
> "print" and "screen" types, to distinguish between embossed and dynamic
> braille:
> 
> media="braille embossed"
> media="braille displayed"

We also had mentioned things like "tty", "hugefont", "lowres".
We need to decide what exactly we want in HTML4 vs being adopted by
market practices.
 
> 2. As Daniel has pointed out, the desire to refer to an image as the
> bullet which is to be displayed by graphical user agents when rendering an
> unordered list, can be adequately met by means of style sheets. I would
> like to add my vote to this approach.

I'm waiting for more input from the CSS gurus to understand what's
missing here.
 
> 3. I also support the proposal to reserve an "abbrevdic" link type;
> indeed, if memory serves correctly, I was responsible for originally
> raising that issue in this forum. Consideration should also be given,
> though perhaps not in the short term, to the relationship between
> abbrevdic and the phonetic markup action item: would it be appropriate to
> encode phonetic information in a dictionary that can be linked to a
> document, rather than embedding it within the document itself? One
> advantage of this strategy is that the same phonetic information can be
> re-used across a range of documents, and it can act as an exception list
> to regulate the pronunciation offered by speech synthesizers.

Agreed. We need to discuss it further.
I've added your comment onto this item's page.

> 5. Returning to the media type issue, Al's suggestion to define style
> properties that would simulate a "text only" environment for the benefit
> of authors who are only able to view their documents with a graphical
> display, should be considered. If it is deemed necessary, then an
> appropriate media type should be reserved. There are already services
> available on the web which can return the document in the form in which it
> would be displayed by Lynx. Whether such services provide an sufficient
> solution or whether a style sheet along these lines is needed, remains an
> open question that Al can perhaps elaborate upon.

I also think it's a useful addition. 'linear text only' might better
convey what I'm thinking of.
 
> 6. I would appreciate clarification of the "alternate SS semantics" issue.
> The messages referred to on Daniel's web page are all on the W3C member
> site, hence inaccessible to non-members.
 
I raised the issue with Bert Bos (W3C staff here in Sophia) and he
posted a first message to the CSS WG mailing list, then I followed up.

Two messages follow: one from Bert and one from me.



From: Bert Bos (bbos@mygale.inria.fr)
Date: Fri, 27 Jun 1997 15:51:25 +0200 (MET DST) 
To: w3c-css-wg@w3.org
Subject: html4-970620: About LINK REL="alternate stylesheet"
-------

Why is there an "alternate" keyword in "rel=alternate stylesheet"?

Can somebody remind me? The only line about it I could find in the
minutes was: "usability concerns, concerns about overloading TITLE
attribute" (see http://www.w3.org/MarkUp/Group/9702/24-html-wg)

I guess that "usability concerns" refers to the fact that the
"alternate" keyword is hard to understand and to use correctly, with
which I agree.

What "overloading" means I don't understand. It must have been the
reason why "alternate" is there now, and it must have been a stronger
reason than "usability"...

Here is a correct example:

(1)
<link href="styleA" rel="stylesheet" >
<link href="styleB1" rel="stylesheet" title="B">
<link href="styleB2" rel="stylesheet" title="B">
<link href="styleC" rel="alternate stylesheet" title="C">
<link href="styleD" rel="alternate stylesheet" title="D">
<meta http-equiv="default-style" content="B">

A is a "persistent" style, A+B is the default style, A+C and
A+D are alternative styles.

What is the difference with this one, without "alternate":

(2)
<link href="styleA" rel="stylesheet" >
<link href="styleB1" rel="stylesheet" title="B">
<link href="styleB2" rel="stylesheet" title="B">
<link href="styleC" rel="stylesheet" title="C">
<link href="styleD" rel="stylesheet" title="D">
<meta http-equiv="default-style" content="B">

It is also quite easy to create incorrect examples:

(3*)
<link href="styleA" rel="stylesheet" >
<link href="styleB1" rel="stylesheet" title="B">
<link href="styleB2" rel="stylesheet" title="B">
<link href="styleC" rel="alternate stylesheet" title="C">
<link href="styleD" rel="alternate stylesheet" title="D">
<meta http-equiv="default-style" content="C">

Note the META. Is A+C now the default, or is this an error?

(4*)
<link href="styleA" rel="stylesheet" >
<link href="styleB1" rel="stylesheet" title="B">
<link href="styleB2" rel="alternate stylesheet" title="B">
<link href="styleC" rel="alternate stylesheet" title="C">
<link href="styleD" rel="alternate stylesheet" title="D">
<meta http-equiv="default-style" content="B">

Are styleB1 and styleB2 cascaded together? And if so, is B an
alternate or not?

It seems to me that dropping the "alternate" keyword will make things
easier to use, harder to get wrong, and easier to implement.

Bert

-- 


From: danield@w3.org
Date: Mon, 30 Jun 1997 13:32:26 +0200 
To: w3c-css-wg@w3.org
Subject: Re: html4-970620: About LINK REL="alternate stylesheet"
-------

To summarize what I got:
no TITLE: persistent style, always applied (modulo media match)
TITLE: a UI is presented for the end-user for selection by name.

What's not clear is the role of REL="alternate stylesheet" vs
REL="stylesheet" vs the presence of TITLE or not.

- Can a persistent (no TITLE) stylesheet be declared alternate and
what does it mean ? (I guess not)

- If there is a persistent and a non-persistent (non alternate), are
they both applied by default or just the persistent ?

- If there are no alternate in the list of TITLEd stylesheet, which
one is used and what is presented to the user in the UI ? Same
question with and without a persistent SS present.

- If there are only alternate in the list of TITLEd stylesheet, which
one is used ? Same question with and without a persistent SS present.

The example given in the spec only covers the case where one TITLEd
stylesheet is alternate and the other is not, in which case it is
applied by default and the other is presented in the UI. Fine, but
couldn't that be achieved using some order of appearance in the HTML
or the HTTP stream ?

Received on Monday, 15 September 1997 09:48:58 UTC