[S] future features of splus

eric.gibson@pharma.novartis.com
Tue, 31 Mar 1998 23:56:21 -0500


Thanks to Charlie Roosen, Stephen Weller, and Doug Johnson (thanks for the
fax) for their kind responses.

However I need to make more clear the marketing opportunity S-Plus is
missing.

A typical statistician working in the pharmaceutical industry has to
calculate sample sizes
every week of his or her working life. You go through a game with the
clinician as to what
sample size is needed for his delta, then he changes his delta (because he
doesn't have the budget)
and wants another sample size, and so on and so forth. The end result is a
number of power curves
sketched together by using nQuery or PASS. (Keep in mind the sample sizes
are usually needed
with only an hour or two's notice, so there often isn't time to write
programs or download a library
of functions written by someone you have never met etc. etc.)

IMAGINE WHAT COULD BE DONE WITH SPLUS AND THE TRELLIS GRAPHICS ENGINE!
The power curves for various sample sizes, deltas, standard deviations, etc
could be easily
printed out in a color graphic that any clinician would love to see and
have no trouble interpreting.

Alas, most pharmaceutical companies are only beginning to discover SPlus.
Don't get me wrong, every company has a handful of users, but it absolutely
does not
have widespread popularity. Many senior statisticians in the
pharmaceutical industry
are diehard mainframe sas users who have only heard of Splus and have no
idea
what it is.

nQuery is not that great, PASS is not that great, SAS has little to offer,
etc. etc.
The door is wide open if only SPlus will step through. I am told that in
version 4.5
there will be power features for one and two sample tests of means and
proportions.

Well, that's a start, and I will certainly take it. But....
Get out your green Graybill and program a suite of functions for
linear models!
I have my own crude functions that get me by. But good power
functions built into Splus should be a must. I like the functionality
of being able to program anything I want in SPlus, but one should not have
to program everything!

And once some of the sas diehards hear that Splus has the best power
functions as well as the best survival functions and the best graphics
functions
(the latter two of which are already true), they might finally come over.

Kind Regards,
Eric Gibson Ph.D.
Novartis Pharmaceuticals

-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news