Comments below..

On Thu, 03 Feb 2005 17:27:33 +1100, Eric Scheid
<[EMAIL PROTECTED]> wrote:
> 
> On 3/2/05 5:09 PM, "James Snell" <[EMAIL PROTECTED]> wrote:
> 
> > What is the model for archiving with Atom?  One or more distinct Atom
> > feeds that each contain a collection of old entries? If so, what we
> > need then is a just container for feeds.  One that encapsulates by
> > reference rather than by direct containment.  Perhaps a top level
> > document like the following would be sufficient?)
> >
> > <!-- archive-index.xml -->
> > <index>
> >  <archive href="2001/archive-feed.atom">
> >     <title>Archive Title</title>
> >  </archive>
> 
> Your example format doesn't include dates for when the referenced items were
> last updated, nor unique IDs for disambiguation, etc etc ... why not
> instead:
> 
> <feed>
>     <head>
>         <title>The 2001 Archive of My Big Fat Fenugreek Blog</title>
>         [...]
>     </head>
>     <entry>
>         <title>January 2001 Archive</title>
>         <summary>208 posts over 19 days</summary>
>         <link href="..." type="application/atom+xml" rel="alternate" />
>         [...]
>     </entry>
>     <entry>
>         <title>February 2001 Archive</title>
>         <summary>311 posts over 23 days</summary>
>         <link href="..." type="application/atom+xml" rel="alternate" />
>         [...]
>     </entry>
> </feed>
> 

This is better.  I guess I just hadn't grok'd the idea of using
entries to reference feeds, but now that I see the angle brackets I
get it.  (perhaps I just need some coffee).  As long as it is legal
for an "alternate" link to reference a feed document, this works
perfectly.   I like the containment-by-reference approach.  Best of
all, this could be accomplished without requiring any changes to the
existing core..

> 
> > Just a single new @rel value for the link element in an atom feed head
> > can then be created that points to this archive index document.
> >
> > <feed>
> >   <head>
> >       <link rel="archive" type="application/atom+xml" href="..." />
> >   </head>
> >   [...]
> > </feed>
> 
> this is exactly what I had in mind with PaceCollections.
> 

Well, from your Pace: "Atom Feed Documents are properly collections,
of which "feeds" are just one semantic. Other semantics for Atom
Collection Documents include "archives", "directory", "comments",
"trackbacks", "pings", "parts", "versions", and so on."

If we define a "feed" as a collection of entries, archives,
directories, comments, trackbacks, pings, parts, versions, and so
forth, then the name "feed" is descriptive enough to get the point
across.  I'm not sure I follow the connection between PaceCollection
and the archive @rel attribute.  Did I miss something?

> e.
> 
> 


-- 
- James Snell
  http://www.snellspace.com
  [EMAIL PROTECTED]

Reply via email to