It is interesting to see people rediscover old ideas. Robert Scoble has a post entitled Silicon Valley got my attention: the future of Web businesses where he writes

What is Zvents capturing? Well, they know you like football. They know you probably are in San Francisco this weekend. And, if you click on one or two of the events, they know you’re interested in them. Now, what if you see an ad for a pair of Nikon binoculars. If you click on that, then Zvents would be able to capture that as well.

Now, what other kinds of things might football fans, who are interested in binoculars, who are in San Francisco, want to do this weekend? Hmmm, Amazon sure knows how to figure that kind of problem out, right? (Ever buy a Harry Potter book on Amazon? They suggest other books for you to buy based on past customer behavior!!!)

It goes further. Let’s say this is 2007. Let’s say that Google (or Yahoo or MSN) has a calendar "branding" gadget out. Let’s say they have a video "monetization" gadget out. Zvents could build the calendar "branding" gadget into their page. What would they get out of that? Lots of great PR, and a Google (or MSN or Yahoo) logo in everyone’s face. But, they would also know where you’d be this weekend. Why? Cause you would have added the 49ers football game to your calendar. So, they would know where you are gonna be on Sunday. And, that you just bought binoculars. Over time Google/MSN/Yahoo would be able to learn even more about you and bring you even more ads. How?
...
It’s all attention. So, now, what if Zvents and Google shared their attention with everyone through an API. Now, let’s say I start a new Web business. Let’s call it "Scoble’s tickets and travel." You come to my site to book a trip to London, let’s say. Well, now, what do I know about you? I know you were in San Francisco, that you like coffee, that you just bought some binoculars, that you like football. So, now I can suggest hotels near Starbucks and I can suggest places where you’ll be able to use your binoculars (like, say, that big wheel that’s in the middle of London). Even the football angle might come in handy. Imagine I made a deal with the local soccer team. Wouldn’t it be useful to put on my page "49ers fans get $10 off European football tickets."

Four years ago, while interning at Microsoft, I saw a demo about Hailstorm which made me suspct the project's days were numbered. The demo involved a scenario very similar to what Robert describes in his post. Just substitute Zvents with "online CD retailer" and calendar gadget with "upcoming concerts gadget" and that Robert's scenario was the Hailstorm demo I saw.

The obvious problem with this "Attention API" and Hailstorm is that it requires a massive database of customer behavior. At the time, Microsoft's pitch with Hailstorm was that its online retailers and other potential Hailstorm partners should give all this juicy customer data to Microsoft then pay to access it. It doesn't take a rocket scientist to tell that most of them told Microsoft to take a long walk of a short pier.

Now let's use a more concrete example, like Amazon. The folks at Amazon know exactly what kind of movies, music and books I like. It's possible to imagine them making a deal with TicketMaster to show me upcoming concerts I may be interested in when I visit their site. The reverse is also possible, Amazon may be able to do a better job of recommending music to me based on concerts I have attended whose tickets I purchased via TicketMaster.

What's the first problem that you have to solve when trying to implement this? identity. How do you connect my account on Amazon with my account on TicketMaster in a transparent manner? This is one of the reasons why Passport was such a big part of the Hailstorm vision. It was how Microsoft planned to solve the identity problem which was key to making a number of the Hailstorm scenarios work. Almost half a decade later, the identity problem is still not solved.

Identity is just problem #1.

If you scratch at this problem a little, you also will likely find an ontology problem as well. How do I map the concepts in Amazon's database (hip hop CDs) with related concepts in TicketMaster's database (rap concerts)? The Hailstorm solution was to skip solving this because it was all coming out of the same database. However even simple things like mapping Rap to HipHop or Puff Daddy to P. Diddy can be fraught with problems if both databases weren't created by the exact same organization. Trying to scale this across different business partners is a big problem and is pretty much a cottage industry in the enterprise world.

Thus Ontologies is problem #2.

There are more problems to discover as one attempts to build the Attention API and an Attention economy. At least it was a fun trip down memory lane remembering my intern days. :)