The submitted link forgot the 'l' in cloud. Here's the actual link: http://www.techcrunch.com/2009/09/09/rsscloud-vs-pubsubhubbu...

The article itself is worth a read. The only thing that bugs me is that the author, and a lot of other people covering this story, don't mention that RSSCloud doesn't support Atom at all (or more accurately it does in theory but not in practice). I think that's pretty significant.

Crap - don't know how that happened. I've reposted the article with the correct link if you care to thumb that one up instead.

The article mentions that the explicit unsubscription of PuSH is an advantage over the implicit approach of rssCloud. With garbage collection being a major problem of publish/subscribe in distributed systems, the author needs to mention how PuSH deals with it in order to assert superiority. Without acknowledging it, the credibility of the article is limited

That's an important point. You can see how PuSH handles it here: http://pubsubhubbub.googlecode.com/svn/trunk/pubsubhubbub-co...

The coining word PuSH of PubSubHubBub is just ... marvelous

That Bub part of PuSH seems as random bits. But I have got it used in:

  * PubBub for a publishing client;
  * HubBub for a hub; and
  * SubBub for a subscribing client.

I should think the reduced latency from receiving the data within the update would be a greater benefit of a fat ping.