The real time Web?

Much has been said about friendfeed‘s latest UI redesign and how it enables a real time view of content from across the Web. Is it really real time? I mean, content is pulled periodically from other applications into friendfeed so that it can be displayed to the end user.

This post was triggered in part by a tweet from Ian Mikutel:

Read your preso on Activity Streams & Context. Does new FriendFeed with Real Time everywhere ruin your “middleman” argument now?

Ian is talking about slide 12 from my presentation “Activity Streams and Contexts” prepared for Google’s first MiniBarcamp on 4/2/09 — note that this was actually before friendfeed launched the redesign — in Lisbon, Portugal (original here):

I claim that a data propagation architecture based on a middleman (friendfeed, for instance) pulling information from different Web applications is better in terms of scaling than letting everyone pull data from each other without any type of agreement. I also say that it can't be real time, because it needs to obtain data periodically from different end points, thus wasting time on that process.

So, where is the real time Web? Is this approach that friendfeed's presenting us the best we can do? I think we can do much better.