W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > July to September 1999

Re: Content negotiation example needed.

From: Al Gilman <asgilman@iamdigex.net>
Date: Sun, 19 Sep 1999 23:02:39 -0400
Message-Id: <Version.32.19990919123938.03b7b960@pop.iamdigex.net>
To: Nir Dagan <nir@nirdagan.com>, w3c-wai-gl@w3.org
At 11:21 AM 9/19/99 -0400, Nir Dagan wrote:
>One should be careful about the type of negotiation.
>
>For example most browsers DO send an accept- header that corresponds
>correctly 
>to the images that they can handle internally, so negotiating GIF and PNG 
>can be done and will work.
>

AG:

Yes.  But the problem is, the critical distinction here is in the domain
negotiated about, not the manner of conducting the negotiation.  The
difference between GIF and PNG at the user interface is nearly
imperceptible and the information added by moving from one to the other is
miniscule.  So it's fair to consider that only the tool cares and the user
can be oblivious without fear of contradiction.  When the difference is one
of natural language choice, there is often value added by having the user
access several if they are multilingual especially in the case where the
author and reader have different orders of language preference.  Here it is
essential that the user be fully informed and have the ultimate decision.

>We actually need to do a more carful research in what browsers support and
how
>concerning content negotiation and http.

Not just HTTP and browsers.  The leaders in implementing negotiation are
internet print and fax functions where HTTP is not a major player.  The Web
needs to be prepared to be a follower where the shoe fits.

Al

>
>Regards,
>Nir.
>
>At 09:37 AM 9/19/99 -0400, Al Gilman wrote:
>>I expect Max is right.  Unfortunately, this would mean Tim was wrong.  This
>>wrinkle is in the Guidelines because nobody wanted to argue the point with
>>Tim Berners-Lee as we were rushing toward Proposed Rec.  HTTP content
>>negotiation as it works now is probably too subterranean and magical to
>>keep the user sufficiently informed on what is going on and what
>>alternatives there are when there are choices being made.
>>
>>How to maintain the benefits and eliminate the glitches is work for the new
>>CC/PP activity.
>>
>>Al
>>
>>Reference:
>>
>>  Comments from Tim Berners-Lee about Web Content Guidelines
>>
>>  http://lists.w3.org/Archives/Public/w3c-wai-gl/1999JanMar/0403.html
>>
>>Prior correspondence:
>>
>>At 05:24 PM 9/19/99 +0900, Masafumi NAKANE wrote:
>>>My prefence is not to rely totally upon the HTTP content negotiation
>>>when users can choose from multiple languages/content types.  Even
>>>when the server supports the content negotiation and the document is
>>>provided in a way that can make use of the content negotiation
>>>mechanism, it would be better to have links pointing to different
>>>versions of the file.  The reasons I prefer this are:
>>>
>>>1. As far as I know, there are not too many people who have set their
>>>   browsers to make use of this mechanism.
>>>
>>>2. Browsers used in an shared environment may not be configured in a
>>>   way that certain user would do.
>>>
>>>In both cases, users would have no idea if the file is available in
>>>different language if there were no links to other versions included
>>>in the page.
>>>
>>>If browsers have easier-to-use interface to configure the language
>>>preferences, the case 1 above can be easily resolved, though.

>>>
>>>    Cheers,
>>>Max
>>>
>>>At Tue, 14 Sep 1999 13:52:13 -0400,
>>>Chuck Letourneau <cpl@starlingweb.com> wrote:
>>>> 
>>>> "1) Instead of including links such as "Here is the French version of
this
>>>> document", use content negotiation so that the French version is
served to
>>>> clients requesting French versions of documents."
>>>> 
>>>> Ok... after reading all the responses and viewing the examples, I can
>still
>>>> barely imagine that this technique is a "Page Author" responsibility
>unless
>>>> it could be  interpreted to mean:
>>>> 
>>>> [start proposed wording]
>>>> If you create more than one language version or format of a page:
>>>> a) ensure that your Web server supports content negotiation, then 
>>>> b) depending on the requirements of your server, include the appropriate
>>>> markup or name the various files appropriately.  
>>>> See your server's documentation or contact your ISP for further help.
>>>> [end proposed wording]
>>>> 
>>>> Some of the example files I looked at modify the file name like this:
>>>>  .../filename.html.xx (where xx= nl, en, fr, de, sv, ja, etc.), and some
>>>> use <HTML lang="xx"> while some don't.  
>>>> 
>>>> This issue reminds me that there was once a suggestion that the
>>>> Guidelines/Techniques documents have a separate section for HTTP/Server
>>>> Accessibility checkpoints.
>>>> 
>>>> Comments?
>>>> 
>>>> Regards,
>>>> Chuck
>>>> 
>>>> 
>>>> 
>>>> At 12/09/99 02:47 PM , Al Gilman wrote:
>>>> >At 11:45 AM 9/12/99 -0400, Chuck Letourneau wrote:
>>>> >>Thanks to everybody who responded to my request for information about
>>
>>>> >>content negotiation.  I am now going to read through it carefully
>>>> >>(including the external references provided) and try to synthesize
>>>> >>something simple for the curriculum.  I will post my result to the
>>list for
>>>> >>your consideration.
>>>> >>
>>>> >>Regards,
>>>> >>Chuck Letourneau
>>>> >>
>>>> >
>>>> >A good starting point is Koen Holtzman's page on the subject at
>>>> >
>>>> > Transparent Content Negotiation in HTTP home page
>>>> > http://www.gewis.win.tue.nl/~koen/conneg/
>>>> >
>>>> >Which I found with the aid of the search:
>>>> >
>>>> >
>>>>
>>>http://www.infind.com/infind/infind.exe?query=%22content+negotiation%22tim
>e=7
>>>> >
>>>> 
>>>> ----
>>>> Starling Access Services
>>>>  "Access A World Of Possibility"
>>>>   e-mail: info@starlingweb.com
>>>>    URL: http://www.starlingweb.com
>>>>     Phone: 613-820-2272  FAX: 613-820-6983
>>>> 
>>> 
>> 
>===================================
>Nir Dagan
>Assistant Professor of Economics
>Brown University 
>Providence, RI
>USA
>
>http://www.nirdagan.com
>mailto:nir@nirdagan.com
>tel:+1-401-863-2145
> 
Received on Sunday, 19 September 1999 23:02:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:47:00 GMT