You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
For simple grammars (and maybe for all grammars?) its seems possible to determine in advance, if they will lead to infinite recursion, on average. Infiinte recusion happens if, on average the likelihood of picking an arity-three (or larger) section exceeds the likelihood of picking an arity-one section. That is, an arity-n section increases the number of open connectors by (n-2) on average, while an arity-one link reduces the number of open connectors by one. Since each section is weighted with the probability of it being chosen, it is a "simple matter" of computing the expectation values.
Its not entirely simple, when there are multiple connector types, but ... it seems feasible to do in code, and can provide advance warning of infinite recursion, before generation is even started. Seems low-priority, for now. See the export-to-gml.scm example for an explicit calculation for the simple case.
The text was updated successfully, but these errors were encountered:
For simple grammars (and maybe for all grammars?) its seems possible to determine in advance, if they will lead to infinite recursion, on average. Infiinte recusion happens if, on average the likelihood of picking an arity-three (or larger) section exceeds the likelihood of picking an arity-one section. That is, an arity-n section increases the number of open connectors by (n-2) on average, while an arity-one link reduces the number of open connectors by one. Since each section is weighted with the probability of it being chosen, it is a "simple matter" of computing the expectation values.
Its not entirely simple, when there are multiple connector types, but ... it seems feasible to do in code, and can provide advance warning of infinite recursion, before generation is even started. Seems low-priority, for now. See the
export-to-gml.scm
example for an explicit calculation for the simple case.The text was updated successfully, but these errors were encountered: