[R] Powell's unconstrained derivative-free nonlinear least squares routine, VA05AD

Berwin A Turlach berwin at maths.uwa.edu.au
Thu Jan 19 02:55:00 CET 2006

G'day David,

>>>>> "DK" == David Kinniburgh <dgk at bgs.ac.uk> writes:

    DK> I have used Mike Powell's optimization routine (VA05AD) from
    DK> the Harwell Subroutine Library (HSL) for more than 20 years.
    DK> [...]

    DK> Now that I have converted to R, I will miss my trusted
    DK> friend. I have started using nls() but have not accumulated
    DK> enough experience to compare the two. It would be great if
    DK> VA05AD could be an option in there (algorithm="Powell"). To
    DK> this end, I recently enquired of the custodians of the HSL
    DK> whether it would be possible to make it freely available to
    DK> the R community. The answer was basically 'yes'.  [...]
You better ask again. :-)

I presume VA05AD is in what is now called the HSL archive?  If it is
part of HSL 2004, then I don't see a way how it could be incorporated
into R given the information on their web site.  

Even for the HSL archive, it is stated at
http://hsl.rl.ac.uk/archive/hslarchive.html that

        "[...] HSL Archive codes are now available without charge to
        anyone, so long as they are not then incorporated into a
        commercial product; the latter is, of course, still permitted
        subject to a commercial licence. [...]"

which sounds as if it would be possible to interface (parts of) HSL
archive to R.  But then that URL goes on ans states:

        "Access to the Archive is by means of a short-lived individual
        password-controlled account. Potential users are asked for
        brief details of the use they intend to make of the package(s)
        they aim to download. Users are also asked to accept a
        conditions-of-use form, and are not permitted to divulge their
        userid and password to anyone else, nor to distribute any
        codes they obtain to a third party." 

IANAL, but as I understand the GPL, the last part of that sentence
will make it pretty impossible to incorporate into R (or distribute as
an R package) (parts of) the HSL archive routines.



========================== Full address ============================
Berwin A Turlach                      Tel.: +61 (8) 6488 3338 (secr)   
School of Mathematics and Statistics        +61 (8) 6488 3383 (self)      
The University of Western Australia   FAX : +61 (8) 6488 1028
35 Stirling Highway                   
Crawley WA 6009                e-mail: berwin at maths.uwa.edu.au
Australia                        http://www.maths.uwa.edu.au/~berwin

More information about the R-help mailing list