Date: Mon, 16 Jun 2008 09:29:29 -0700
Reply-To: Stephen McDaniel <stephen@STEPHENMCDANIEL.US>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Stephen McDaniel <stephen@STEPHENMCDANIEL.US>
Subject: Re: OT Friday Humor
In-Reply-To: <000c01c8cd88$0449c950$832fa8c0@HP82083701405>
Content-Type: text/plain; charset=ISO-8859-1
I will say something controversial here... SAS as a language is "dead."
Even though I use it everyday and I love it, I don't think very many new
"programmers/developers" choose it as their career or language of choice,
they instead happen upon it for a project (usually in grad school) and pick
it up, often finding lucrative niche opportunities in corporate areas
desiring such skills. Statisticians at most schools are biased towards R or
Stata for various reasons (cost and openness.) People in Social Science
areas are still leaning towards SPSS.
R + Perl/Python + My SQL can do many of the things SAS can and are freely
available, also R is object oriented and in some ways more flexible. That
said, SAS still has many advantages in the corporate world, but one big
disadvantage- price. It is imperative that SAS continue pushing an
effective analytic platform into colleges and universities, even if it means
(gasp!) SAS would actually lose money in this market by just giving the
software to students, etc.
If I was a low cost start-up I would lean towards R + Perl/Python + My SQL
for statistical analysis and towards SQL Server + Excel + Tableau if it
was more of the business analyst crowd. And then there is JMP, which
totally undercuts SAS on pricing, especially for lower and mid-scale data
mining needs. In fact, I just recommended JMP to a friend due to pricing,
even though SAS EM is a fuller and more complete tool set- they just can't
justify the 10-20X price differential. It is sort of strange that SAS
undercuts itself in data mining, but I am glad that they do or I would need
to point people elsewhere in all but the biggest data mining projects...
For large teams in larger companies, it is hard to beat SAS, but that is
less true today than ten years ago... Hopefully SAS skills in 2020 won't be
like COBOL skills in 2002... I think this is largely up to SAS to determine
as their is still plenty of opportunity to capture.
Stephen
On Fri, Jun 13, 2008 at 12:02 PM, Mary <mlhoward@avalon.net> wrote:
> And then there's this wonderful note:
>
> Tonight's Iowa Eye Banquet has been canceled, for obvious reasons.
>
>
>
> -Mary
>
| 
