Given the list of the reaction step rules, the reactions computed with the program aim at representing the elements of a chemical intuitive class of organic reactions (for example the reactions with only substitutions or eliminations under certain conditions). The computation is called complete if the computed reactions represent all reactions in the chemical intuitive class. It is called correct if every computed reaction represents a reaction in the intuitive class.

Generic reaction step rules which have a high embedding quote (like for example 4 points where all 6 edges between them and all points are wildcards that match everything) will determine a complete and incorrect representation, for which the performance is reduced by the large number of generated formulas.

A few specific rules, those with low embedding quote, will produce an incomplete, correct representation possibly with a good performance.

A high number of specific rules can be taken for a complete, correct representation, but in this case the performance is reduced by the amount of time needed to apply all the rules.

Performance can be significantly improved by the use of the options

1. RULE APPLICATION BOUND, which tells the program to stop applying a rule when the given number of products has been generated with that rule,
2. DOMAIN, which limits the domain of points between which bonds can change, or
3. DISTINCT, which lowers the number of generated formulas (but slightly increases the checking activity at any generated formula).

The 2-step option TWO STEPS is also faster but it misses the odd levels. Therefore the best is often a series of computations with and without TWO STEPS.

<< back