Today I stumbled on PubSub.com, a system that
you filter over one million weblogs and information streams to find
the content you're looking for, in real time. It's like searching
If I understand correctly, they idea is that you can define search parameters to filter out aggregated news items that you might be interested in, presented as a personalised RSS feed. The difference with other weblog news services being that they search in newly aggregated items, instead of some history of past news items. Also, it is supposed to be usable with any kind structured data.
But where does the PubSub from their name come in, then? Well, apparantly, this service is only live for a few weeks now, and they have just added an experimental RESTian interface to notify you of new messages. They call your webserver with a HTML POST request containing the result of a match to your search query.
Bob Wyman weighs in on why PubSub.com is different from *, but I think it doesn't go far enough. As I see it, they like to be a hub for all this data, and I don't know if a centralized service is the way to go here. Like I said earlier I envision a distributed system with news authors publishing news (or any other data) themselves, to which people can subscribe, and having this data spread through hierarchies of pubsub repeaters.
Having such a distributed system gives power to both the publisher
as to the subscriber. For example, the publisher needn't be concerned
with having their items move slower than others because of commercial
interests of the centralized service, and the subscriber can sure that
he only receives data he has asked for (no extra stuff added for your
Still, I'm curious where PubSub.com will go next. It looks like a cool service with potential. I'm holding out for the support of Jabber PubSub.