Foreword
My freshman year was a notable moment in our school's history. The summer before, a group of rising seniors won the Superquest competition, collecting as their prize an ETA-10P: ours was the first high school with a supercomputer. The following spring, this journal published its inaugural edition. The thrill and ambition of scientific research was in the air, though as a nervy freshman merely attempting survival, I admit that my goals were rather less grand.
A scientific late bloomer compared to many of my TJ peers, I just wanted to write code to solve interesting problems. At the time I didn’t appreciate that such curiosity could be called research, nor could I have envisioned how the coming revolution in data acquisition (whose common analogies to disasters like avalanches and tsunamis should perhaps give us pause) would transform our approach to science. By coupling high-performance computing with unprecedented data streams, mathematical and statistical models grow ever more vital to the scientific enterprise. It is hard to find a discipline that has not felt this influence. My love of coding eventually matured into research in computational linear algebra, and I delight in the ways matrix algorithms empower data science, enabling insight across so many domains.
Yet given more data than we can handle, it is tempting to presume that we can capture reality by adding another term or two to our model, by a more sophisticated regression, by a deeper neural network. Among the most important scientific lessons I learned in college came from an English professor, who taught me that “models are metaphors." This same point is made by René Magritte's painting The Treachery of Images, and by Alfred Korzybski’s observation that “the map is not the territory.” Models provide a vital perspective on reality, allowing us to comprehend, analyze, simplify, even optimize. But we must not mistake the model for reality. While this is true throughout science, this fact becomes ever more essential when we model the behavior of one another. People are not numbers. The data scientist must embrace complexity, humbly and honestly acknowledge a model's limitations, and be wary that the innumerate often place unreasonable faith in models they are unable to question. For this latter reason data literacy is vital for all citizens, especially leaders and policy makers.
Many are drawn to science as a force for doing good at remarkable scale, and our historic moment provides ample evidence, widely and joyously celebrated. Yet we must also acknowledge that science is not a universal good: history reveals too many cases where, whether through accident or malice, science has taken humanity to places we would rather not have traveled. If we want to do good, how do we know what makes good science? Can you imagine turning down work because you do not believe it is sound? What anchors you? Just as we need numerate leadership, so too must our researchers learn lessons from the humanities, appreciating (as our colleagues are quick to point out) that science is not the only route to wisdom.
Among my intellectual heroes is Lewis Fry Richardson, a pioneer of computational weather forecasting and mathematical models of international conflict, a man whose career was profoundly shaped by his high principles. His landmark book, Weather Prediction by Numerical Processes (drafted while serving as an ambulance driver in France during World War I) includes a remarkable vision for parallel computing: a spherical theater mimicking the planet, staffed by human computers who pass their calculations to their neighbors as they collectively integrate the partial differential equations modeling the weather. His "fantasy" ends with this coda: "Outside are playing fields, houses, mountains and lakes, for it is thought that those who compute the weather should breathe of it freely.” His lesson is a good one for burgeoning scientists: to lead a principled life, tackling problems of interest and import and learning from others, all while remembering to enjoy the world beyond your models.
Congratulations, authors, on your remarkable work! We celebrate your achievements, as we hope and trust that the bright sparks contained in this issue of Teknos kindle into magnificent light. May that work create a world better for us all to inhabit, and may you also find ample time to delight in it.
Mark Embree
Professor of Mathematics, Virginia Tech
TJHSST Class of 1992