“Writing code helps me to understand the development process so that I can make better judgments about it”: An interview with SAS founder John Sall

Authors: Ajay Ohri and Sunakshi Bhatia

John Sall is a legendary figure in statistical computing. He not only co-founded SAS Institute, the leading enterprise analytics and statistical computing provider for the past three decades, but he also created the major statistical software, JMP (originally named from John’s Macintosh Program).

Sall designed, developed and documented many of the earliest analytical procedures for Base SAS® software and was the initial author of SAS/ETS® software and SAS/IML®. He also led the R&D effort that produced SAS/OR®, SAS/QC® and Version 6 of Base SAS.

In the late 1980s, Sall noticed a niche that SAS software was not fulfilling. Researchers and engineers – whose jobs didn’t revolve solely around statistical analysis – needed an easy-to-use and affordable stats program. A new software product, today known as JMP®, was launched in 1989 to dynamically link statistical analysis with the graphical capabilities of Macintosh computers.

Now running on Windows and Macintosh, JMP continues to play an important role in modeling processes across industries as a desktop data visualization tool. It also provides a visual interface to SAS in an expanding line of solutions that includes SAS Visual BI and SAS Visual Data Discovery. Sall remains the lead architect for JMP and he was also elected a Fellow of the American Statistical Association in 1998.

In this interview, John Sall talks of the things that continue to drive him, his favourite statistical method and why he loves coding.

 

1. In your 2009 interview with Decisionstats, you said “making the commitment to evolve led to an amazing sequence of growth that is still going on after 35 years.” How have JMP and SAS Institute evolved in the past five years?

Our customers have much larger data sets now. SAS has responded with architectures for high-performance computing across many servers. JMP is still confined to single machines, but has been tuned to handle larger problems much faster. In the past five years, JMP has also been further strengthened in the areas of design of experiments (DOE), reliability and modeling. And we’ve made strides in consumer and market research.

2. What is your favourite statistical method? What do you like about it?

One of my favourites is optimal design of experiments because it enables you to learn the most from a given number of runs of an experiment. This is an area in which Bradley Jones is one of the pioneers. Brad is Principal Research Fellow at JMP and leads development of the DOE platform in JMP. In particular, Brad has enriched the field with breakthroughs in I-Optimal split plot, definitive screening and supersaturated designs. I’m a huge fan.

3. SAS now has SAS University Edition, which can be either downloaded or tried out on Amazon AWS Marketplace. What are your plans for JMP in the cloud?

The cloud is a growing option for organizations to obtain value from their data through analytics, and SAS is very much moving toward the cloud. For JMP, we are watching technological change, but we have no plans at the moment for JMP in the cloud. JMP works great on the desktop. Check out my blog post about why the desktop isn’t dead.

4. SAS Institute has an amazing suite of products. How do you make sure that products don’t end up either competing with themselves, cannibalizing sales or adding to brand confusion or clutter?

Part of my job is to keep JMP focused on a different customer group than SAS. JMP is designed for scientists, engineers and other researchers. SAS customers are largely in business and IT. Of course, there’s going to be some overlap. But we tend not to compete because our products are very different, and we market our products to different groups.

One of my favourites is optimal design of experiments because it enables you to learn the most from a given number of runs of an experiment. This is an area in which Bradley Jones is one of the pioneers. Brad is Principal Research Fellow at JMP and leads development of the DOE platform in JMP. In particular, Brad has enriched the field with breakthroughs in I-Optimal split plot, definitive screening and supersaturated designs. I’m a huge fan.

5. Some people may not be aware of what JMP does. Could you just take us through the products JMP offers, what sectors they are aimed at and how they complement other languages like SAS and R?

JMP is point-and-click data analysis software designed for scientists, engineers and other researchers. It has a rich visual interface that makes it easy to get to discoveries and insights. JMP Pro includes everything in JMP plus advanced features. JMP Clinical and JMP Genomics are combinations of JMP and SAS for clinical and genomics researchers. And we have an iPad app that lets you explore data on the go. JMP integrates with SAS, R and MATLAB. Integration with R lets you use features in R with the point-and-click interface of JMP.

6. Do you still write code? Describe what keeps you motivated in coding.

I do still write code. It’s something I enjoy, and I’m good at it. I like to express myself in code. Writing code helps me to understand the development process so that I can make better judgments about it.

7. Women tend to be underrepresented in statistical computing in 2015. Please comment or critique this statement and some initiatives you have taken as an organization to improve the situation.

SAS VP of Advanced Analytics R&D Radhika Kulkarni is a woman, and so are a number of SAS department heads. Half of the development managers at JMP are women, as are many JMP statisticians and testers. To encourage more women to pursue tech careers, SAS has a summer program for students called R3 that recruits women studying computer science, computer engineering, computer information systems, management information systems and statistics. Many of those students end up working at SAS upon graduation.

Ajay Ohri is founder of analytics company DecisionStats and author of two books on statistical computing. He can be contacted on LinkedIn.