[Rd] Optimisation methods

Kurt Hornik Kurt.Hornik@ci.tuwien.ac.at
Fri, 1 Dec 2000 09:05:03 +0100

>>>>> Robert King writes:

> On Thu, 30 Nov 2000, Prof Brian D Ripley wrote:
>> On Fri, 1 Dec 2000 Robert.King@mailbox.gu.edu.au wrote:
>> > I don't want to re-invent the wheel, and I'm trying to code up something
>> > that does a Nelder-Mead simplex method to minimise a non-linear objective
>> > function.  (I'm porting something I originally wrote in matlab, using the 
>> > optimisation toolbox funciton fmins).
>> optim(method="Nelder-Mead")
> Thanks.  Perhaps I should just learn to spell like the OED :-) 

>> Now how did this not come up in a search: perhaps because of the way you
>> spell optimization?  I have
>> which shows we probably need to find a way to check variant spellings.
>> (BTW: just in case anyone tells me the British spell with an `ise', the
>> Oxford English Dictionary does not.)

> I guess this means I should change the spelling of the
> parameterisation in the gld package functions.  What do people think
> of adding extra arguments to functions to take account of different
> spellings like this:

> example.fn _ function(parameterisation="fing",parameterization=NULL)
> {
> if (is.null(parameterization)) {parameterization_parameterisation}
> else {parameterisation_parameterization}
> print(parameterisation)
> }

> Is there a better way?

Well, first of all use `<-' rather than `_'.

We've discussed the spelling issue on r-core in some detail.  You may
notice that e.g. there are both license() and licence().  [There is a
more important issue in gray versus grey, I guess.]

I personally think that there is nothing that says that argument names
must have any meaning in any language and hence one may spell them in
any way we want.  But more realistically, in the above case I'd rather

	exampleFn <- function(param = "fing", ...)

which will make everyone happy.

The real issue I see is how people can find documentation if they use
different spelling, and I think we need a fuzzy grep which allows near
matches for that purpose.

r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch