W3C home > Mailing lists > Public > whatwg@whatwg.org > November 2007

[whatwg] Feed autodiscovery draft may be resurrected

From: James M Snell <jasnell@gmail.com>
Date: Mon, 05 Nov 2007 11:08:09 -0800
Message-ID: <472F6A19.6010309@gmail.com>
The type attribute indicates the kind of document being referenced,

  <link rel="service" type="application/atomsvc+xml" href="..." />

Because some clients (like windows live writer) support multiple
protocols, we add a class="preferred" to our service link to indicate
which link the server prefers the client to use [1],

  <link rel="service" class="preferred"
        type="application/atomsvc+xml" href="..." />

Beyond that it's completely undefined... which, of course, is not a good
thing.  We need a standard or at least documented-and-commonly-used
solution to this problem.


- James

Ian Hickson wrote:
> On Mon, 5 Nov 2007, James M Snell wrote:
>> Just as an example, Windows Live Writer uses the service link to 
>> discover the location of the Atompub service document; which it then 
>> uses to configure itself to interact with one or more Atompub 
>> collections.
>> For instance, in IBM's internal blogging environment, any employee can 
>> have one or more blogs.  When they use WLW, rather than manually 
>> configuring the client for each individual blog, they can simply point 
>> WLW to what we call the "dashboard" at http://blogs.tap.ibm.com/weblogs. 
>> WLW will get that page, look for the service link, get the service 
>> document, and automatically pull out the blogs that user can edit.
> Ok, but how does it know that it's an Atompub service document, as opposed 
> to a FooBar service API? Is the "type" attribute required, or...? (And if 
> so, does this mean that you can only have one type per api, and one api 
> per type?)
Received on Monday, 5 November 2007 11:08:09 UTC

This archive was generated by hypermail 2.3.1 : Monday, 13 April 2015 23:08:38 UTC