Psychologists know that - however much insights do occur - learning is often a gradual process, one of familiarization, of stumbling into puzzles and resolving them by proposing simple hypotheses in which a new problem is seen as like others already understood and performing experiments to test the latest "theory."
Computer-based microworlds can be seen as sets of programs designed to provide virtual, streamlined experiences, play worlds with agents and processes one can get to know and understand. Properly designed microworlds embody a lucid representation of the major objects and relations of some domain of experience as understood by experts in the area. This is where the knowledge of the culture is made available, in the very terms in which the microworld is defined.
Children can absorb that knowledge because the microworld is focussed not on problems to be done, but on "neat phenomena" - these show the power made available by knowledge about the domain. If there are neat phenomena, then the challenge to the knowledgeable expert is to formulate so crisp a presentation of the elements of the domain that even a child can grasp its essence. The value of the computer is in building the simplest model which an expert can imagine as an acceptable entry point to his own richer knowledge.
If there are no neat phenomena that a child can appreciate, he can make no use of knowledge of the domain. He should not be expected to learn about it until he is personally engaged with other tasks which will make the specific knowledge worth learning as an aid in achieving some other personal objective.
Publication notes: