At 9/6/2001 at 07:53 AM, Prof Brian D Ripley wrote:
John Chambers has said that it is time for some S standardization effort,
and I back that. Concepts like data frames should be frozen for all
time. It will help the users, who may be using two or three S
implementations simultaneously.
A vote of agreement with Prof Ripley.
As a computer language matures, concepts must be standardized. It is
clearly possible to add features without revising or breaking what already
exists. (A notable example is the evolution from Fortran 77 to Fortran
90. The language added extensive functionality, yet code that *was*
standard conforming still *is* standard conforming.) The commitment to
stability is what distinguishes a serious computer language. Part of what
makes complex code worth writing is expecting it to work tomorrow, and in
ten years, and in twenty years, with little or no change. That doesn't work
for macro languages in commercial products, or for many proprietary
development facilities, but it should hold for serious languages. I would
hope to see S in the "serious language" category, and I hope the
implementers (whether they be commercial or noncommercial) would agree.
--
Michael Prager, Ph.D. <Mike.Prager@noaa.gov>
NOAA Center for Coastal Fisheries and Habitat Research
Beaufort, North Carolina 28516
http://shrimp.ccfhrb.noaa.gov/~mprager/
|