[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Syndication of updates and deletions of content



Hi all,

A very short introduction of myself: my name is Pier and I've been somehow involved with Apache for the past 7 years... I currently work for VNU Business Publications in London, and being in the publishing market, we're starting to look at different ways to use syndication protocols throughout our systems.

I'd like to start by saying that (for the moment) I am not talking about Atom's Publishing Protocol, but rather of the Syndication format, which (for us) is at the moment more pressing.

The one thing I find difficult when reading throughout the Syndication Format specification is whether it allows the syndication of "deleted" resources.

Let's start with a very very very simple example: I have an article <http://www.vnunet.com/news/1157398> about Windows XP SP2, and what can happen on this resource is that it can either be published or deleted.

Published is easy, yesterday there was no resource, today we have one. In Atom, I would syndicate this as:

<?xml version="1.0" encoding="utf-8"?>
<feed version="..." xmlns="...">
<head>
<title>VNUNET.COM Feed</title>
<link rel="alternate" type="text/html" href="http://www.vnunet.com/atom/index.atom"/>
<modified>2004-08-17T16:23.52Z</modified>
<author>
<name>VNU Business Publications Limited</name>
</author>
</head>
<entry>
<title>Microsoft lists apps affected by XP SP2</title>
<id>http://www.vnunet.com/news/1157398</id>
<link rel="alternate" href="http://www.vnunet.com/news/1157398"/>
<author><name>Iain Thomson</name></author>
<issued>2004-08-17T08:29:29Z</issued>
<modified>2004-08-17T08:29:29Z</modified>
</entry>
<entry>
....
</entry>
</feed>


This means that a resource has been published. In this format I can also specify that a resource was modified, if (for example) the modified date specified in the new feed is different from the original feed in which it first appeared.

I cover, basically, insertions and modifications of resources with no problems.

Now, how would I syndicate the fact that a resource is gone away? Of course I won't include in the syndication feed all of our content (more than 100k articles), but only what happened in the (let's say) past 24 hours. But how can a subscriber to my Atom feed know whether a resource is simply not included, or it was actually removed?

I thought about having an "entry" structured in this way:

     <entry>
       <title>Microsoft lists apps affected by XP SP2</title>
       <id>http://www.vnunet.com/news/1157398</id>
       <link rel="alternate" href="http://www.vnunet.com/news/1157398"/>
       <deleted>2004-08-17T08:29:29Z</deleted>
     </entry>

but I'm not sure whether this makes sense... Probably it might be better not to include the <title> tag, as it wouldn't make sense anymore (what's the title of a non-existant resource), but I'm just shooting out ideas.

What do you guys think? Is there already a way to do this?

Thanks to you all...

Pier

Attachment: smime.p7s
Description: S/MIME cryptographic signature