The newspaper, that daily chronicle of human events, is undergoing the most momentous transformation in its centuries-old history. Online versions are proliferating, attracting young readers, and generally carving out a sizable swath of the news business. In the United States alone, 34 million people have made a daily habit of reading an online newspaper.

An online news site can change minute by minute and generate a different front page for each reader. The most interesting and useful customization involves capturing information about the readers' interests from their past behavior. There's already a model for that--the recommendation systems used by Web sites like Amazon, TiVo, and Netflix.

Using information on past purchases, movie ratings, or items viewed, these systems steer consumers to items from among the thousands or millions they have on offer. How long before newspapers, or even other media sites like those mentioned above, borrow this idea for news?

Greg Linden works at Microsoft Live Labs. From 1997 to 2002, he worked at, first writing its recommendation engine and then leading the software team that developed the company's personalization systems.

In an article in the March issue of IEEE Spectrum, Linden talks about how, researchers at Google and elsewhere have tackled these problems, which must be solved if newspapers are to play the same role in the next century they have in the two previous ones.

It may seem a small step from recommending products to recommending information, but, in fact, doing so is actually quite complex. Stand at the entrance of a Wal-Mart or look at Amazon's home page and the shiny world of their wares seems limitless - but it's not. It is firmly bounded by the constraints of time and warehouse space.

A sprawling Wal-Mart store typically has about 100 000 items; carries a few million. The world of information, on the other hand, is measured in billions of pages and petabytes of data.

Processing data on this scale can require a supercomputer-scale infrastructure well beyond the means of a city newspaper. Recommender systems also face what is known as the "cold start" problem, which stems from the difficulty of rating any item that has not yet attracted the notice of qualified recommenders.

But supercomputer-sized infrastructures are already in place in many purchasing systems online. The cold start problem is also easy to resolve just by making a guess and using big leaps in convergence from there.

So it's not just traditional newspapers at risk from the demands of a busy consumer base. Online newspapers may have a tough road ahead unless they transform into more niche news or increased personalization.

Amazonews? Newsflix? It could happen.