abdera-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dan Diephouse <dan.diepho...@mulesource.com>
Subject Re: Modeling Collections of Collections
Date Tue, 25 Sep 2007 18:31:39 GMT
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
  <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Brian Moseley wrote:
<blockquote
 cite="mid:29a761a00709251105n192f26b8g2160d8e2704d7091@mail.gmail.com"
 type="cite">
  <pre wrap="">On 9/25/07, James M Snell <a class="moz-txt-link-rfc2396E" href="mailto:jasnell@gmail.com">&lt;jasnell@gmail.com&gt;</a>
wrote:
  </pre>
  <blockquote type="cite">
    <pre wrap="">No, but that's a very application specific thing to do; if you're sure
others will be able to figure out what you're doing, then it's fine.
    </pre>
  </blockquote>
  <pre wrap=""><!---->
so is nesting a collection inside an entry. either way, the writer of
a client will have to be told that the feed includes something
non-standard. dan's suggestion makes it explicit.
  </pre>
</blockquote>
Well this is part of my problem with the whole "RESTful services don't
need a description language thing". It'd be very nice from a machine 2
machine point of view if I could have some metadata inside my
&lt;collection&gt; which goes beyond &lt;accept&gt;, because for
anything beyond well understood content types like html, images, etc,
you're stuck.&nbsp; For instance, I may know that my collection accepts
"application/atom+xml;type=entry", but:<br>
<br>
1.How do I know what types of XML content my &lt;entry&gt; might hold.
I.e. does it contain a &lt;customerInfo&gt; document of the type
application/vnd.acme.customer+xml? It seems that if you aren't using
media entries, you have to have the accept set to
"application/atom+xml;type=entry", no? <br>
1a. Should &lt;customerInfo&gt; be included as a direct child of
&lt;entry&gt; or &lt;content&gt;?<br>
1b. If the later is valid, I can set a contentType on &lt;content&gt;.
Then could I also do
&lt;accept&gt;application/vnd.acme.customer+xml&lt;/accept?<br>
<br>
2. How do I relate the mime type to the schema so that the other
machine can browse to it?<br>
<br>
Ideally I'd like to see something like:<br>
<br>
&lt;workspace&gt;<br>
&lt;collection&gt;<br>
&nbsp; &lt;description mimeType="application/vnd.acme.customer+xml"
descriptionType=<a class="moz-txt-link-rfc2396E" href="http://foo.bar/relaxNg">"http://foo.bar/relaxNg"</a>
href="blah.rng"/&gt;<br>
&lt;/collection&gt;<br>
&lt;/workspace&gt;<br>
<br>
<blockquote
 cite="mid:29a761a00709251105n192f26b8g2160d8e2704d7091@mail.gmail.com"
 type="cite">
  <pre wrap="">dan, what about a rel or name attribute on the collections to
distinguish them apart?

&lt;atom:entry&gt;
...
&lt;app:collection rel="purchaseOrders" .../&gt;
&lt;app:collection rel="contacts" .../
&lt;/atom:entry&gt;
  </pre>
</blockquote>
That could work. I didn't realize that was legal! (although legal !=
standard ;-))<br>
<br>
- Dan<br>
<pre class="moz-signature" cols="72">-- 
Dan Diephouse
MuleSource
<a class="moz-txt-link-freetext" href="http://mulesource.com">http://mulesource.com</a>
| <a class="moz-txt-link-freetext" href="http://netzooid.com/blog">http://netzooid.com/blog</a></pre>
</body>
</html>

Mime
View raw message