First of all, they are monitoring 6.5 million blogs and seeing 600,000 posts a day. That’s pretty amazing. However, many of the 6.5 million blogs have feeds that haven’t pinged them or updated in some time, so they prefer the more conservative 3,606,000 active blogs as the number they actually monitor day-to-day.
They’ve also changed the interface, making it far easier for users to see that PubSub is really about defining a search for subscribing in my aggregator (those searches can actually include several searches, for example: napsterization.org/stories — to get all instances of someone linking to me, combined with “mary hodder” and “napsterization” to get all instances of the keyword use in posts, and then being able to define interesting subsets of the blogosphere.. so that I only get posts say, in the 40% to 80% conversational middle of blogs as they are ranked by inbound links… and on and on…).
PubSub does something interesting that the other blog services don’t do… which is that they don’t collect the data and store it… instead they match the data to the subscription searches and other things they collect and monitor, and then toss it. So they don’t have to worry about structuring it in a database for later. It’s an interesting premise, this idea of not searching historically but rather searching just what comes through, which they call “prospective search.”
Bob Wyman of PubSub adds:
Mary Hodder points out on her blog that the PubSub matching engine doesn’t “collect the data and store it” like search engines do. This is, of course, because we focus on prospective search (“Searching the Future”) rather than the retrospective search that search engines provide. Since we’re searching the future, it doesn’t make sense to build up a collection of data that we’ve seen in the past. The only thing we care about is what’s new. We’re not just “yet-another-search-engine”… We’re working on the “other half of the search problem” that hasn’t gotten much attention in the past.
Since we don’t store data, we can’t let you know what has been said about any particular subject in the past. We can only promise to let you know if what you’re interested in is mentioned in the future. Thus, a prospective search service only handles half of your search needs. The other half must be provided by using a retrospective search engine like Google, Feedster or Snap. Prospective Search compliments Retrospective Search rather than replaces it. The two are best used together.
This distinction between prospective and retrospective search maps exactly to the different phases of “research” that people typically pass through when seeking information. For instance, imagine the you just came across a reference to PubSub.com and you decided that you wanted to find out more about us. First, you would probably do a retrospective search. You would use Google or Snap to find out “What is known?” about us — what has been written in the past. Then, if you decided we were interesting, you might think to yourself: “Let me know whenever there is something new about PubSub.” That second question would be a “prospective search” and it is what we do at PubSub.com.