🚀 go-pugleaf

stetzer@csd4.csd.uwm.edu (Frank Stetzer) writes:
>A faculty member here is interested in the NCSS statistics package
>for PC's.  I don't have any experience with NCSS; does anyone have
>a feel for how it compares to Systat, SPSS/PC+, etc., in terms
>of functionality, performance, stability, etc? 

I'd say it's a functional, stable system.  I'm not sure I'd compare it to
the big-time stat packages, but it's pretty capable and has some nice 
graphics capability.  For the price, it's a great deal.  I've used it for
several years.

It's especially nice for the occasional user - it uses screen forms for
setting up inputs and the like.  There's not a lot of syntax to try
and remember.  OTOH, that approach gets frustrating for expert users.

>I take it that NCSS is no relation to CSS.  Is that in fact so?

No connection that I'm aware of.  The author is a statistician somewhere
in Utah.  The whole program is a set of linked programs, (I believe) in
compiled BASIC.  No matter.  The accuracy seems pretty good.  In order to
overcome the limitations of memory availability, it reads data line-by-line
from disk, so performance is tied to disk I/O (which you can speed up by
stashing the data files on a RAMdisk).

I don't hesitate to recommend the package, based on my experience, especially
if your're working on a short budget.

--
----------------------------------------------------------------------------
-- Archie Warnock                   Internet:  warnock@hypatia.gsfc.nasa.gov
-- Hughes STX                       "Unix --- JCL For The 90s"
-- NASA/GSFC