W3C home > Mailing lists > Public > w3c-wai-ua@w3.org > January to March 2001

Re: problems with Checkpoint 2.1

From: Al Gilman <asgilman@iamdigex.net>
Date: Thu, 01 Feb 2001 14:21:50 -0500
Message-Id: <200102011910.OAA1958029@smtp2.mail.iamworld.net>
To: Jon Gunderson <jongund@ux1.cso.uiuc.edu>, w3c-wai-ua@w3.org
At 12:48 PM 2001-02-01 -0600, Jon Gunderson wrote:
>Responses in JRG:
>  An example would be an SVG graphic used in a web page.  The SVG
>>actually containts a lot of words, and the words are internally stored in a
>>legible form following the suggestions of the access note for SVG.  But the
>>User Agent doesn't implement SVG.  The user should be able to glance at the
>>source as text and find the natural language embedded there even 'though
>>don't get the graphic rendition.  In addition to this, if the user has a
>>for the format of any object in the content, then the user must have the
>>to have that content rendered with that player.
>JRG: So this would require source views for unsupported or unrecognized 
>formats.  Is XML based formats a special case?

XML is not alone, but it is an important example because it is where there are
going to be lots of resources where the UA hasn't implemented the dialect yet,
but a source view would be of help.

The definition I suggested had SGML and "Anything recognized as of Internet
Media Type text/*" as well as XML.

In pratcial terms, if the UA's text viewer will accept it, it should be up to
the user then to determine that it is useless garbage or contains intelligible
information.  Text formats are any format where the text viewer is a feasible

>>Another application scenario is that if what the author of the SMIL thought
>>user didn't need is an audio track, then the user must have access to an
>>rendering of that object if the User Agent supports that type of audio
>>at all.  The user should be able to chose from the full range of feasible
>>options, based on the formatting of the data and the formats implemented by
>>The user must have their choice including a) if it is a format that the User
>>Agent implements, the user must have the choice of formatted rendering
and b)
>>if the format is a text format and the User Agent implements Internet Media
>>Type text/plain, then the user must have the choice of text presentation. 
>>requirement is the union of these two, not the intersection.
>>Another example:  If the document links to a stylesheet in text/css , and
>>User Agent implements text/plain, then the user must be able to view the
>>of the stylesheet whether or not the User Agent implements CSS.
>Are you saying that the UA must provide source views of linked resources?
>Is an image SRC considered link information?
>Do you know of anything that does this?

This comes in two parts.  On the Web, you can't trust the format of a link. 
There may be a type indication on the link, but it's 'advisory.'

So, if there's a linked resource, and you suport hyperlinking, then you owe
user to recover the resource, if asked.

Once the resource has been recovered, you have an object in hand.  This value
of the recovered object is what has a type.  If the type is one you can play,
then you owe the user the chance to play it.  The ability to play it in some
context or by itself will depend on a) the rendering rules of the format where
the link was and b) the limitations of the media that the user has the User
Agent displaying into.  So where the rendered value of the recovered object
shows up depends on lots of things.  That's not the point of this discussion. 
What I am talking about is the steps to "play it again, Sam."

Experiment:  In a web page, set the SRC of an IMG to point at an SVG file. 
Then, in Lynx, turn the "image links" feature on to be sure you get recovery
capability.  You may have to adjust the mime.types database to cast image/svg
to text/plain.  But I believe you can configure Lynx as it ships today to walk
through the steps I discussed above.

In Windows, this is handled by the OS.  Any referenced object of an
unrecognized type is immediately turned over to an "OpenWith" dialog in which
the user is queried as to what tool to handle this file with.  So that's lots
more than two Aps that provide this facility.

Or maybe we need to look at the mime.types architecture that is supported by
Lynx and lots of Unix mailers as a second implementation of this case.

Actually, I should have referred to the 'External' capability in Lynx, where
the user can interactively select the player for an object where there is a
link to the object in the page.

I believe that Netscape also has menu extension capabilities like this.


>><possible new>
>>    2.1 Make all [158]content available through the user interface. For 
>> formats
>>that the User Agent implements (see [162]checkpoint 6.2), make the content
>>available through the implemented rendering process   In addition, for all
>>ref]text formats, make a source text view available. [Priority 1]
>JRG: should "implemented" be "specified".  If the rendering is not 
>specified anywhere, do we want to include the concept of using conventions?
>>[edit to reduce the note, perhaps.]
>>           Note: In general, user agents render content through the user
>>           interface according to the instructions of implemented
>>           specifications (see [162]checkpoint 6.2). However, some users
>>           may not be able to access this content, even when the user
>>           agent renders it according to specification (e.g., for scripts
>>           or style sheets). Therefore, this checkpoint includes the
>>           requirement that user agents make available raw text content
>>           (including attributes, style sheets, etc.) through the user
>>           interface. Although this view (which may be a [163]document
>>           source view) is an important part of a solution for providing
>>           access to content, it is not a sufficient solution on its own
>>           for all content. This checkpoint does not require that all
>>           content be available through every [164]viewport. See
>>           [165]guideline 5 for more information about programmatic access
>>           to content.
>>           [166]Techniques for checkpoint 2.1
>>AG::  Insert definition at end of [new ref]
>>Text formats:  This includes the following two subcategoies.  First,all
>>objects given an Internet Media Type of 'text', that is text/plain,
>>text/* or text/anything.  Second, all SGML and XML applications
regardless of
>>Internet Media Type, including HTML 4.01, XHTML 1.1, SMIL, SVG, etc.
>Jon Gunderson, Ph.D., ATP
>Coordinator of Assistive Communication and Information Technology
>Division of Rehabilitation - Education Services
>College of Applied Life Studies
>University of Illinois at Urbana/Champaign
>1207 S. Oak Street, Champaign, IL  61820
>Voice: (217) 244-5870
>Fax: (217) 333-0248
>E-mail: jongund@uiuc.edu
>WWW: <http://www.staff.uiuc.edu/~jongund>http://www.staff.uiuc.edu/~jongund
>WWW: <http://www.w3.org/wai/ua>http://www.w3.org/wai/ua
Received on Thursday, 1 February 2001 14:10:24 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:49:29 UTC