The two graphs above show the underlying reality behind one of the major shortcomings of the World Wide Web and a shortcoming that a digital library for SMET education can directly address. Click anyplace inside the border above. The graph on the left is a graph of all things technological -- the speed of the average desktop or notebook computer, the number of Web sites, and the speed of the average Internet connection. The graph on the right is a graph of all things human -- my personal reading speed, the number of hours in the day, and the number of days in the week. Because our own time is limited, we cannot personally examine everything that is available -- even at the local bookstore, let alone on the Web. One particularly compelling metaphor describes using the Web as trying to drink from a fire hydrant using a straw.

We need mechanisms that will help human users find high quality resources that match their needs. Like an analog library, a digital library employs two basic kinds of mechanisms -- selection of the materials acquired for its collections and search and reviewing mechanisms to help users find and choose material in the collections.

As presently conceived, a digital library for SMET education would rely on its member collections for selection with different collections having different selection criteria. The library would, however, offer two very powerful kinds of mechanisms to help users find high quality and appropriate resources.

• Multidimensional search and prioritizing mechanisms.
• Selective portals.

Multidimensional search and prioritizing mechanisms

The library would employ multidimensional search mechanisms that would select and prioritize hits based on several different sources of information including the following:

• Information within the body of the resources -- for example, occurences of key words or other textual information. Search engines might even look at visual or auditory content.
• Metadata attached to each resource. Metadata can be a very powerful tool. It can be used, for example, to find resources that are appropriate for a particular level or experiments that can be performed with certain equipment. By itself, however, its usefulness may be limited since it is usually supplied by the developer or by a cataloger working with the original resource.
• Third party reviews including reviews by professional societies and by users. Each object in the library will be identified by a digital object identifier that will enable search engines to find reviews of the object.
• Citations. The search engine Googol, for example, is able to supply surprizingly useful prioritized lists of hits by looking at information about the linking among sites. Very roughly, if a site is cited often then it is more likely to be worth looking at.
• User-supplied information. Searchers could optionally supply information that would help search engines choose and prioritize hits. For example, information about available software and laboratory equipment.

Selective portals

Because portals can be designed for specific users and in some cases might serve only members, they might play a particularly effective role in helping users find the best resources for their purposes. For example, portals might make more use of "amazon.com" kinds of information. The on-line bookseller amazon.com posts comments submitted electronically by customers and also information based on customer orders -- "other readers who purchased this book also purchased ... "