That’s a good point, and I enjoy how you compare it to combination products in medicine!

Do you think SAS is doing something wrong that causes JMP to be slow in statistical innovation, or is it perhaps the nature of a business to lag a bit behind?

Does a bleeding-edge procedure gain credibility or stature simply by being added to a big-name product like SAS? If so, what responsibility do companies like SAS, Mathematica, SPSS, etc. have to evaluate the worth and applicability of new methods? Is it enough that the market seems to demand things?

I always go back to 3D pie charts. We all know they’re atrocious, but customers want them. So what is the proper role of market demand in the evolution of a product? How prescriptivist should software providers be?

]]>JMP has an excellent user interface but a small progress in providing new statistical routines.

R with its user contributed packages has the most dynamic progress in development of statistical procedures I am aware of.

Therefore I expect great thinks from the combination of both via the new JMP to R interface in JMP 9 and future releases of JMP. ]]>

I think your statement […] sums it up perfectly, and I also enjoyed the use of kerfuffle.

Your post doesn’t mention it (although I’m sure you know) R is not only a Linux tool. R runs perfectly fine in the beloved Windows environment.

The comments section on AnnMarie’s blog are priceless: http://www.thejuliagroup.com/blog/?p=433

I think JMP is making an excellent move by simply allowing its users to connect to R. The kerfuffle arises from SAS/JMP users confused over what do to with R. Simple; you don’t *have* to do anything. I’ve used R quite extensively in graduate school, yet in industry the open source phobia is blatantly apparent (there are several exceptions to this rule – Google comes to mind). I think the strategic move here is to capture R users (and future SAS users) with JMP as the common interface. I see this as a win for R, SAS, and both their user bases.

Posted by Mike Olson, Analysis Manager

EYC

I think your response was well-balanced. I haven’t seen the word, “kerfuffle” in a long time, but using that word surely places the controversy in a certain context. I wasn’t aware of the controversy until your post.

I use JMP and I appreciate all of the support and the webinars and the conferences, although I’ve yet to go to a seminar. And certainly, the link between JMP and SAS helps me with analytical “naming dropping.” Most people have heard of SAS.

I’ve been tempted to learn R or use the R-add-in for Excel for the analyst group I manage. The reasoning is this: 1) We don’t do enough statistical analysis to support multiple JMP licenses. 2) There’s a large learning curve with JMP as well. I’ve been working with JMP for a year, and there’s still plenty of stuff I don’t know how to do. So for the team, another software package to learn. 3) There’s a core set of simple stat routines that we would use: the basic descriptive stats and graphs and some t-tests. Excel is a pain to use, even for a simple histogram. So the R-add-in for Excel is very tempting. Just haven’t tried it yet.

Posted by Richard Giambrone, Manager, Business Intelligence & Analytics

Patterson Companies