Some remarks to the implementation

The implementation does not concern:

the speed of the developed application
The speed of every system based on domain knowledge heavily depends on the quality of used knowledge. Since we did not try to optimize either the quality of knowledge being used or the system performance, the execution time of the solution (PC 386 sx / 25 MHz - 243 sec and PC 586 / 75 MHz - 9 sec) has only illustrative meaning. In spite of this, it indicates that CLIPS requires more powerful computing facilities.
economic use of system resources
Some means of CLIPS were used in a considerably careless way trifling away CLIPS resources (e.g. excessive number of activations in agenda although only one rule was fired and the other activations were removed - for instance when the activity was transferred to the task evaluate for the first time there were 1490 activations on the agenda).
consistent use of CLIPS means
A lot of CLIPS means able to make the final implementation more effective were not used in our test implementation, especially those dedicated to handle complex programs (e.g. the possibility of partitioning the program into several independent modules with the focus only on one module at the same time).

The program was written in an incremental way. First, the CLIPS system version 6.03 in the form of 16 bit application running under WfW 3.11 was used. At a certain stage of the program development the used CLIPS version kept crashing (with the obligatory general fault protection message) and was not able to find a solution. In the next stage the CLIPS system version 6.03 in the form of 32 bit application running under DOS 5.0 has been used.

This experience can indicate that the CLIPS resources are not unlimited and that the consistent use of all means offered by CLIPS to decrease the resource consumption is necessary, especially for the implementation of complex problems.