by Chris Trimble
Every reader is probably familiar with Mosaic, Gopher, and the many other emerging information viewers for the Internet. Everything from the White House press releases to pictures of Cindy Crawford can be found on the information "web," of which the Internet is a superset. There's so much information on this web that it's almost impossible to keep track of the information popping up daily; therefore, we wind up with central indices and indices to those indices. I propose that we adopt a new strategy for managing information. This technology, proliferating wildly like it is, will provide mainly useless information unless a fundamental idea behind the web changes drasticallyÑits distributed nature.
The original idea of the Web was probably to allow sites to have semi-formatted on-line documentation, some information about their site, and some fun stuff. With newly-found corporate interest in the Internet, this idea has changed drastically. Many feel that the Web should have relevant information such as printed magazines and books in electronic format. Typically, unless this information is at a site that pops up automatically when you run one of these information viewers, you will have a difficult time finding the info youÕre looking for. It becomes a click-and-search techniqueÑclick on a link, find a "dead end" (a page with no links), go back, click on some moreÑuntil the information is found. Obviously, information searches can't stay this random or no one will use this technology. As a result, some have proposed numerous types of indices to the information that is exists on the Internet. While I agree that this is important, it is more important to put this information into several large central servers as well.
Moving the web from a wholly distributed nature to a more centralized design is analogous to changing from a door-to-door search technique to a library visit. In the library, a person can get many of the cross-references with a walk across the room; and if the information isn't available, inter-library loan is usually possible. If one goes door-to-door, there is always the risk of the information not being there: the person may have moved, which is often the case in the world of computers, since computers are constantly changing roles in organizations.
Central servers, because of their considerable computing power and the effort put into creating them, are pretty much guaranteed to stay up on the network. They would be able to easily cross-reference between themselves and to coordinate the entirely disjunct organization we see today. Finally, they can provide a standard interface for users, with intelligent searching programs that can rifle through this conglomeration of dataÑnot search aimlessly through hundreds of sites before finding it.
There are, of course, considerable downfalls to this kind of system, which is why not all non-central servers should be abolished. If there were only three central servers on the entire information netÑowned by Pewlett Hackard, SoftMicro, and Crapple ComputerÑitÕs conceivable that such companies could be owned by a parent company like MotopolaÑthus leaving the entire information network in the hands of one company. So, to allow for freedom of information (information anarchy) and personal Mosaic pages, there should of course be more than these few central servers on the network; I simply feel that the important information that people want must be centralized in a way that will make such an information network efficient and reliable.