A Trainer's Perspective on the JMP Discovery Conference
I can't believe that the conference has come and gone. That milestone means that half of 2008 is over already! Before the rest of the year flies by, I thought that I should add my thoughts to this blog. I had nothing to do with the preparation for the conference this year, so my observations are completely unbiased.
The pre-conference workshops were quite full, even though it was Father's Day. Scripting is a popular subject and seems to draw more interest every year. The Beginner's Scripting Workshop and the Advanced Scripting Panel Discussion were conducted in parallel. I noticed that extra laptop computers were brought in to accommodate the over-flow of students in the beginner's workshop that was taught by Paul Marovich. Some of these students stayed after the conference for more training about scripting.
I helped with the panel discussion. Wayne Levin and Diane Michelson of Predictum, Inc. were the first two panel members to present topics dealing with 'abstraction'. They explained that in many cases, the best or most elegant solution is a script that writes another script. Huh? Instead of evaluating the whole script as is the usual case, part of the script is suspended as an 'expression' that can be modified to suit the particular situation better, before it is evaluated. They backed this claim up with a lot of examples. For my part, I was asked to talk about script variables. I covered all the different kinds of values that can be saved in a variable, the difference between global and local variables, and some best practices for when to use each kind or a literal value instead. We were all busy right up until the bus dragged us back to the conference hotel!
I found many of the conference presentations gave me ideas about training opportunities as well. The passion of the speakers and the eager reactions of the audience is an indication to me that some of these topics may be introduced into our existing courses or demand an entirely new one:
> Damon Stoddard talked about how JMP, in general, and scripting, in particular, were key in rapidly expediting Six Sigma projects in the Microsoft hardware division. Are other users held back from rapid success because they don't have the ability to script JMP to do what they need? We offered the initial scripting course within a year of JSL appearing as a new feature in JMP 4. Should we offer short, quick lessons aimed at specific outcomes like database queries or custom graphics to match the 'just in time' nature of the user problem?
> Susan Dorsey, from the University of Maryland, talked about discoveries in the genetics behind chronic pain. JMP Genomics 3 represents a mature solution that still enjoys aggressive development to keep pace with this field. The developers have conducted many training seminars already. If we were to make regular training courses, what should they look like. Is our style and format for traditional subjects the best way for investigators specializing in genomics research?
> Lise Varner, from Cambrios Technologies, talked about how she has exploited scripting. In less than a year of attending a public class for the Introduction to the JMP Scripting Language course, she had developed several of her own scripts to the point of mastery. She showed how these scripts reduced the tedious and repetitive aspects of her work in reliability and accelerated life testing. All of the scripting instructors would like to take a pat on the back for a success story like Lise's case, but she struggled, too. What can we learn from her case about the choice of the topics covered, teaching methods, examples, and exercises for practice that might have reduced or eliminated her struggle?
> Joni Keith talked about how she has helped Bush Brothers. This company has a long history and many successes to date but is still innovating to bring new products to market. They rely heavily on sensory analysis, a sophisticated science that attempts to collect qualitative information and make it more quantitative. It also relates to John Sall's opening keynote address when he introduced the 'choice designs' and analysis in JMP 8. Will product designers and Web site designers prefer traditional training about design of experiments or training devoted to these kinds of studies?
> Bradley Jones talked about an important new class of fractional factorial designs. They are the same size as the resolution III designs that were originally obtained by applying group theory operations to generating words. The new designs, which are only available from JMP custom design, do not exhibit complete confounding between the main and two-way interaction effects in the model, thus providing more information to an investigator. This discovery is the latest of several JMP developments in recent years to assist screening studies. Should we develop a short course dedicated to the design and analysis of screening experiments?
> Jose Ramirez, from W. L. Gore, talked about several kinds of analyses in JMP. In particular, he started by showing accelerated degradation analysis. This method has been known for some time but is not commonly performed. It is just the thing, though, when you have a stable material and you want to establish its shelf-life but you do not want to wait a long time to see it fail. We hear many questions about this kind of need in the area of drug substance and product expiry. How can you use the data from an accelerated stability test or stress test to determine real-time shelf-life?
There were many other interesting presentations; I highlighted these few because of their implications to training opportunities.
The post-conference training classes were quite full. The Introduction to the JMP Scripting Language ran until the very end of the week and managed to keep almost every student chair occupied. We were also pleased to offer two special workshops, Analysis of Dose-Response Curves and Stability Analysis using JMP. Both of these half-day workshops will be expanded to a full day with the addition of new topics based on student and customer interviews. In addition, specialized tools (scripts) that support the training and the practice of these methods will be an added bonus to students. These courses represent a new direction for JMP training. (Don't worry, our current curriculum will remain intact and up to date.) The new curriculum will be industry-specific and problem-oriented. These workshops addressed common tasks in the pharmaceutical and biotechnology industry. What other tasks or problems does this industry need training for? What other industries have similarily well-defined tasks that use JMP analytical methods?
I guess I better stop here and get to work. We have a lot of new training to ponder and develop, and half of the year is already behind us! You can help. If this entry makes you think of some new training that you would like us to offer, please let us know.