I have experimented with various AI techniques, but most of them were too time-consuming to build and too complex to demonstrate without a lot of pedagogical development. However, one simple one can be adapted to both the nouvelle and the classical AI approaches. This is the production system.
The idea of a production system is that you have two kinds of ``memory'' or storage. One - short-term or working memory - holds the data currently being worked on, as a set of symbols. The other - long-term memory - consists of rules which operate on these symbols, looking for certain patterns and replacing them by other patterns. You can think of them as stimulus-response rules: given a certain stimulus or pattern of symbols, your response is to replace it by a different pattern. Abstractly, this situation is shown below, in figures 1 and 2. The numbers, incidentally, are not part of the rules, and are just there for reference.
Figure Some production-system rules.
Figure Symbols in working memory.
I have not said anything about what these symbols might mean. The production system itself has no knowledge of this, and its operation does not depend at all on their meaning. The symbols only gain their meaning by virtue of their relationship to other symbols, rules, perceptions, and motor commands.