5. Pictures of Knowledge Development and Dissemination

5.2. Generic vs. Context-dependent Encapsulation of Knowledge

One big advantage of the journal system that would not be lost in a move towards review boards is in their context-specific nature. Each journal is able to employ reviewers, editors, review criteria editorial policy and style that is appropriate to the area it has chosen to cover and these can develop to keep pace with the developments in that area. Even the broad generic science journals such as Nature and Science have their known focus areas.

The big weakness of big paper archives and generic search engines is that they have to cover a very broad range of topics written about in a broad range of styles. Effective search strategies require the application of domain-specific meta-information. An analogy can be drawn with the `No Free Lunch' theorems in computer science [5] which show (in a very abstract way and over an unnaturally large set of possible search spaces) that no search algorithm is better than others over all possible search spaces.

In the fields of AI and machine learning context-dependent learning and reasoning are increasingly seen to be powerful tools (e.g. [1]). Further there is increasing evidence that human learning and reasoning is inherently context-dependent (e.g. [6]). Review boards allow the introduction of context-dependency to a greater extent than journals because as well as covering different areas in different ways they can also cover the same papers in different ways for different audiences and purposes.

A Proposal for the Establishment of Review Boards - Bruce Edmonds - 16 MAR 99

[Next] [Previous] [Up] [Top] [Contents]

Generated with CERN WebMaker