Hi.
Referring to
https://issues.apache.org/jira/browse/MATH-1345
I propose the following changes:
1. Make default values "private" (rationale: it's an implementation
detail).
2. Add a new instance variable in "AbstractRealDistribution" for
holding
the solver accuracy (which is currently accessed through a
"protected"
accessor, "getSolverAbsoluteAccuracy()", that returns the value of
"solverAbsoluteAccuracy" field defined in each subclass)
This field is to be initialized at construction ("final").
3. When a subclass overrides the "inverseCumulativeProbability" method,
the base class's solver is not used; in such case the subclass's
constructor will pass "NaN" for the solver's accuracy.
4. Make "getSolverAbsoluteAccuracy()" method "public".
Objections?
Regards,
Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org