R-beta: "Comparison of Mathematical Programs for Analysis"

[Ross Ihaka]

Bill Venables writes:
 > For example R and S+ do lack built-in polygamma functions, but it
 > is trivial to write very effective polygamma functions and
 > suddenly they are available.  (See, for example, the digamma and
 > trigamma functions in the MASS library.)  For other software
 > systems doing the same thing is a major undertaking or even
 > actually impossible unless the vendors do it for you.  That's a
 > major difference that these comparisons usually hide completely.

Just a minor correction - we actually have
	digamma
	trigamma
	tetragamm
	pentagamma
built in to R (someone wanted some of these for fitting glms).  I
think I got bored at the "penta" stage.  We could have a general
"poly" if anyone *really* wants it.  The underlying code is in place.

I have to agree in general that these kinds of comparisons are at best
misleading.  Usually it is only direct experience which tells you
whether a particular bit of software is right for your particular
problem.

One thing the page didn't seem to address was scalability.  Many bits
of software (ours included) work just fine on small problems, but are
pretty useless when a truly large problem comes along.
	Ross
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
r-help 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-help-request at stat.math.ethz.ch
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=