Date: Thu, 28 Apr 2011 11:13:01 -0500
Reply-To: Charlie Huang <charlie.chao.huang@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Charlie Huang <charlie.chao.huang@GMAIL.COM>
Subject: Re: write a book on SAS programming (for SAS programmers,
not just for SAS users )
Content-Type: text/plain; charset=windows-1252
I agree with Murphy. Many SAS programmers are reluctant to use SAS/IML.
When I was in school, every computer with SAS had SAS/IML. I used it
extensively for many research projects. Once I started to work, I
surprisedly found that many sites donít have the SAS/IML module at
all. As the result, my program based on matrix computation cannot work
properly, which makes me feel awkward. I have to recode them by Proc
FCMP, which is available in SAS/BASE since 9.1. This procedure is a
cheap alternative for SAS/IML, even thought coding it takes more
loops/iterations than SAS/IML.
SAS should do more work to popularize SAS/IML. For example, SAS could
make it part of SAS/BASE or SAS/STAT, which decreases the total
On Thu, Apr 28, 2011 at 10:41 AM, Murphy Choy <email@example.com> wrote:
> Hi Rick,
> Apologies if my statement were biased. I am commenting on the basis of my
> experience in my region which is drastically different from the US. I have
> so far came across only 2 companies who are using SAS/IML. Hopefully take up
> rate will increase.
> Below are my comments which are entirely my personal opinion.
> 1) There are many thousands SAS/IML sites and programmers.
> =>Yes. There are many sites and several programmers who I know use SAS/IML.
> =>When I refer not widely used, I meant that probably only 10% or less of
> the existing SAS users base use it. I could be wrong on this as I do not
> have any statistics on this.
> =>Hopefully more people can adopt it because it has great synergy with data
> mining especially with implementation of algorithm.
> 2) SAS Press has for years been wanting someone to author a book on
> SAS/IML because customers frequently come to the SAS Press booth and ask
> for one.
> =>Yes. I have been waiting for it as well. Great to see it out now.
> 3) My book (Statistical Programming with SAS/IML Software) is selling very
> =>Not surprising. Being the only book for SAS/IML as well as probably the
> only statistical programming book with SAS, it has great appeal among
> SAS/IML and algorithm people.
> 4) My IML blog (http://blogs.sas.com/iml) is one of the top-read SAS blogs
> with hundreds of subscribers.
> =>Your blog is one of the few blogs that explain extremely interesting
> programming concepts as well as solution to problems. A refreshing read for
> people who wants to learn more about SAS/IML. I probably read one of the
> least read but equally interesting one by Clark.
> 5) There are more than 230 references to "PROC IML" in the SUGI
> proceedings on www.lexjansen.com.
> =>I hope that we can have more SAS/IML in the upcoming forum. Would be
> interesting to get OR and IML mixed together.
> However, I do agree with Murphy's general statement: a general getting
> started book does not need to devote very many pages to SAS/IML, since it
> is a more advanced topic.
> =>I think we should get some basic matrix manipulation steps in here. In
> fact, a simple bivariate regression example would be excellent.
> Murphy Choy
> Certified Advanced Programmer for SAS V9
> Certified Basic Programmer for SAS V9
> DataShaping Certified SAS Professional