Meanwhile, in its unofficial way, the Internet is transforming the way information is stored. The traditional function of libraries, gathering books for permanent storage or one-at-a-time lending, has been thoroughly confused. Archiving of the on-line world is not centralized. The network distributes memory. There is a kind of self-replication at work, with data employing humans in the effort to spread and reproduce. Web site by Web site, the data seem as frail as skywriting—smoke in the breeze. Brewster Kahle, estimating the average lifetime of a Web page at seventy-five days, created an Internet Archive to capture and store periodic snapshots of almost the entire Web. It saves pages that have been lost or shut down by their owners. It amounts to about eight terabytes of data. (Tera- is trillion; peta- is next.)

Meanwhile, in its unofficial way, the Internet is transforming the way information is stored. The traditional function of libraries, gathering books for permanent storage or one-at-a-time lending, has been thoroughly confused. Archiving of the on-line world is not centralized. The network distributes memory. There is a kind of self-replication at work, with data employing humans in the effort to spread and reproduce. Web site by Web site, the data seem as frail as skywriting—smoke in the breeze. Brewster Kahle, estimating the average lifetime of a Web page at seventy-five days, created an Internet Archive to capture and store periodic snapshots of almost the entire Web. It saves pages that have been lost or shut down by their owners. It amounts to about eight terabytes of data. (Tera- is trillion; peta- is next.)