Applied Research versus Academic Research
This posting by Michael Rys really resonated with me. While I agree with this call to the academic community to do spend more efforts examining current implementations of technologies, I'm reminded of what a mentor of mine from my days in the Financial Services industry told me:
"The essential difference between an Actuary and a Mathematician is that while the Actuary looks at experiences and proposes theories to explain them, the Mathematician looks at theories and proposes experiences to explain them."
It is that difference of applied research versus academic research that set me off. Last night I had the opportunity to speak a good sized gathering of students from Illinois State University. I asked them "How many of you want to be hardcore developers in the future?" Lots of hands went up. "How many of you want to be hardcore DBAs?" One hand. ONE. I later asked some folks individually how much exposure they were getting to database, information and knowledge management systems at an undergrad level. The answers were disappointing -- it didn't sound like the study of such things had progressed significantly in the last fifteen years. One class, Junior or Senior Level and they probably get through as much content as you would in Database Design for Mere Mortals.
The problem I gather is that there are not good bridges between the academic and industry communities in these domains. The fundamental challenge, it seems, is how do you teach these topics in a meaningful way and from experienced basis in a traditional academic setting? There seems to be an irreconcilable difference between the two, since the academic would say that the implementation shouldn't matter. Problem is -- in the world -- it does. Normalization is great example of that. Folks like Codd and Date have long preached separation of the logical data model from the physical data model. But in practice, the tools and products are heavily geared towards the physical model instead. Worse, the development of products is heavily driven by market forces, not theoretical correctness.
So, do I have a solution? Frankly: No. But I do think were just starting to see the adoption of technologies like XQuery and the Semantic Web which could be such a bridge. I'll call on Microsoft to help enable the Academic community to embrace their efforts and I'll call on the Academic community to take them up on their efforts. A workable middle ground should be able to come out of such efforts, I believe.