Re: [OctDev] econometrics functions

[Michael C. <mic...@ua...>]

On Friday 22 October 2004 12:19, Paul Kienzle wrote:
> On Oct 22, 2004, at 9:49 AM, Michael Creel wrote:
> > Hi again Paul,
> > One thing, I was wondering if it is good policy to add directories the
> > way I
> > did with "econometrics". I think that the stuff I added is different
> > enough
> > and interdependent enough so that it should not just be dropped into
> > the
> > main/statistics directory. I would have used a subdir inside
> > statistics,
> > except I understand that 2nd level subdir of main don't make it into
> > the
> > package. IS that correct? Any comments about using a new dir?
> > Michael
>
> The choice of what goes into a separate directory should be driven by
> the likely audience.  MLE should therefore go into the statistics
> directory
> since it is applicable far beyond the econometrics community.  If there
> are functions specific to econometrics, they should go into a separate
> econometrics directory.
>
> Since everything is in a flat namespace I don't find a hierarchical
> organization within directories particularly compelling, and certainly
> not
> compelling enough to support it in the test and install scripts.  I
> don't
> object to others doing so if they are so inclined.
>
> - Paul

Well, since GMM estimation is also known as "minimum chi square" estimation, 
which is just as venerable as MLE, maybe they should all go there. Once I 
have them properly documented with usage help I'll move them.
Michael