[comp.misc] Wanted SAS vs S comparison

urjlew@ecsvax.UUCP (Rostyk Lewyckyj) (09/10/87)

  [/]
 
This is being posted for a colleague, Dr. Lee Pedersen of the department 
of Chemistry, here at the University of North Carolina in Chapel Hill.
Please reply directly to him, or to me and I will forward the information.
---------------------------------------------------------

UNC-CH is considering purchasing a minisuper computer that
runs UNIX.  S is an interactive statistics package developed
at Bell Labs.  We are most interested in learning its plusses
and minusses in comparison with SAS.
  
We are very interested in finding

a)  people who have had experience with both SAS and S
b)  people who know where there is information (reviews)
    comparing SAS and S
c)  people in a university setting that have used S.


Lee Pedersen   ULGPKG@TUCC.bitnet
          tel. (919)-962-4403

-----------------------------------------------
  Reply-To:  Rostyslaw Jarema Lewyckyj
             urjlew@ecsvax.UUCP ,  urjlew@tucc.bitnet
       or    urjlew@tucc.tucc.edu    (ARPA,SURA,NSF etc. internet)
       tel.  (919)-962-9107

jackg@hpisoa2.UUCP (09/23/87)

As regards SAS and S comparisons:

I used both SAS and S when I was a graduate student at UC Berkeley.

S and SAS have different, almost counterposed philosophies, from
which follow their relative strengths and weaknesses.

S aims at maximizing flexibility.  It's interactive, so if you
need a p-value for a standard distribution, you can get it
immediately (rather than submitting a job).  It is fairly
easily extensible in two ways:  first, by writing macros that
use S's built-in functions; second, by actually writing new
functions for S.  The first method is quicker, but less powerful,
since the macros have to be interpreted into S functions.
Since S provides a fairly limited number of statistical routines,
the ability to extend it is very important.
Finally, S provides powerful graphic capabilities, which are
particularly useful in conjunction with methods like Tukey's
Exploratory Data Analysis.

SAS, in contrast, provides a huge and powerful assortment of
statistical tools.  It may well have everything you need, done
in the way you want it done.  It is not interactive, and it
is not easily extended.

Jack Gerson

bruce@dolqci.UUCP (Bruce Limber) (09/25/87)

>SAS . . . is not interactive.

Correction:  SAS runs in either interactive or batch mode.

-- 
Bruce Limber (NEW ADDRESS:  uunet!vrdxhq!dolqci!bruce)    (202) 535-0640

If we are not careful, we are liable to wind up where we are headed.
                                           (Chinese proverb)

emigh@ncsugn.ncsu.edu (Ted H. Emigh) (09/26/87)

>>SAS . . . is not interactive.

>Correction:  SAS runs in either interactive or batch mode.

Clarification:  SAS is only partially interactive (at best).  If nothing
else, the cost to the user is MUCH larger for the interactive mode
(on our machine with our cost algorithms).
-- 
Ted H. Emigh, Dept. Genetics and Statistics, NCSU, Raleigh, NC
uucp:	mcnc!ncsuvx!ncsugn!emigh	internet:  emigh%ncsugn.ncsu.edu
BITNET: NEMIGH@TUCC                  @ncsuvx.ncsu.edu:emigh@ncsugn.ncsu.edu