Closed kliegl closed 11 years ago
In some sense this is already there: ?lmerControl
says
...: additional arguments to be passed to the nonlinear optimizer (see ‘Nelder_Mead’, ‘bobyqa’). In particular, both ‘Nelder_Mead’ and ‘bobyqa’ use ‘maxfun’ to specify the maximum number of function evaluations they will try before giving up - in contrast to ‘optim’ and ‘optimx’-wrapped optimizers, which use ‘maxit’.
We could go a little farther in being more explicit, e.g. adding an explicit statement of what the maximum value is, but I'm reluctant do too much cross-copying of parameters across help pages: this is 'documentation replication' creep, and if we do too much we'll end up copying all of the information to all of the web pages to enhance its visibility (and ending up with a mess).
I guess you are right. I did not make it down to the "..." when I read the ?lmerControl, but got trapped in the first paragraph about the optimizers. Thanks and sorry.
I entirely agree with Ben.. Even more, as the buzzword 'maxfun' is indeed visible on the help page. In such a case, I always search forward (Emacs C-s; but all other R IDEs have some such capabilities) on the help page for the exaxt term, here 'maxfun'.
Would it be useful to list maxfun=10000 under the ?lmerControl() arguments?