A Thoughtful SIAM Community Weighs In---Students, Journals, SoftwareJune 21, 2011
Letters to the Editor
To the Editor:
I found it quite fitting that an issue (March 2011) focusing on complexity (as the theme of Math Awareness Month, and in Philip J. Davis's review of Complexity: A Guided Tour) also has a wonderful article on the complexity of balancing work and family life ("I Am Not Your Supermom---Balancing Family and Work," by Elebeoba May). While an academic career may be the shortest path to our scientific aspirations, it is rarely the path of least resistance. As a postdoctoral scientist continuing his nomadic quest to find an academic position, I am encouraged to hear that the fears of changing goals, priorities, and motivators are not unique to my own thoughts. As the author puts it, "a little bit of disequilibrium and unbalanced equations" in one's life can be frightening, but also enriching.
Courant Institute, New York University
To the Editor:
I would like to comment on the impact our SIAM student chapter has had at the University of Arizona. Due to generous support from SIAM and our Program in Applied Mathematics, we have been able to host a wide variety of activities in recent years. Examples include research talks from faculty in many areas of science, professional development activities, such as career panels and introductions to computing, and an astronomy evening in the desert. We have also organized math-themed days for elementary students through the Tucson Kids' Club. These events have provided us with additional perspectives on the benefits and opportunities in studying applied mathematics, complementing our professional training as graduate students. Personally, it has also been very rewarding to interact with faculty who are passionate about integrating mathematics into their research. Many thanks to SIAM for its commitment to encouraging and engaging with students in applied math and related disciplines. I highly recommend that every university start a student chapter if it hasn't done so already!
President, SIAM Student Chapter
University of Arizona
To the Editor:
How can we better engage undergraduates in mathematics?
At UMass Dartmouth, we are addressing this issue with an undergraduate student workshop on research in scientific computing, involving students in real mathematical problems that a professional mathematician would be glad to solve. These research problems come from our own research and scholarship in applied mathematics and mathematics education; contrived problems, with the flavor of book exercises, are often a turn-off for students.
Workshop students are eager to learn new mathematical and computational skills: linear algebra, because the matrices they encounter in radial basis function applications are ill-conditioned; GPU and parallel programming, because of computational work on black holes; developing fast algorithms, because of clique-finding problems in biological networks, for example.
Workshop students learn techniques, including Fourier analysis, and skills like LaTeX, Matlab, Python, Mathematica, and R programming faster and more deeply when topics are "just-in-time."
Engaged students are willing, often keen, to discuss their work in seminars, and at conferences, including those of SIAM, MAA, Wolfram Technology, and Sigma Xi.
Alumni of our workshops report that the skills learned assist them in graduate school and in gaining employment.
For some students, research engagement really seems to work, while for others it's less successful. Should we aim for research engagement to permeate the undergraduate mathematics curriculum, or should it just be an option?
Gary Davis and Sigal Gottlieb
University of Massachusetts Dartmouth
To the Editor:
As an early-to-mid-career numerical analyst of PDE, I am concerned that software development is not taken seriously enough as an integral part of progress in applied mathematics. I, and I imagine many SIAM News readers, spend a big fraction of our time building and testing codes as part of algorithm research, often with a larger user community in mind. The impact of high-quality, reliable, documented software can hardly be overstated---one need only contemplate the duplicated human effort saved by the universal adoption of LINPACK/LAPACK. However, there appears to be a certain intellectual and funding-agency snobbery, that this is not "real mathematics." If one "focuses on the software side rather than on the mathematical theory," this is a bad thing (actually, I quote here from the response to an NSF DMS Computational Mathematics grant proposal!). Yet, without shared documented libraries, we are doomed to reinvent the wheel and, worse, to be unable to verify that everyone's claimed wheels indeed roll, or to compare various wheels and adopt only the one that rolls fastest. We need more support for head-to-head tests of algorithms, and "reproducible research," in the words of David Donoho . (Thus, in recent publications I include example codes that let anyone easily test the algorithms within a toolbox environment.) How can we make our culture and funding more supportive of this essential aspect of computational mathematics and science?
 D. Donoho, A. Maleki, M. Shahram, V. Stodden, and I. Ur-Rahman, 15 years of reproducible research in computational harmonic analysis, 2008, http://www-stat.stanford.edu/~donoho/Reports/.
To the Editor:
The article on publication times ("Data Mining: 100 Feet of SIAM Journals," SIAM News, March 2011) presented a clear question: Are publication times in SIAM journals too long? In short, yes. Just this afternoon I was discussing alternative publication venues with a colleague because "SISC will take two years"---about three months of work left, one month of writing, and an estimated 16 months to publication.
By the time this paper would be published, I'd be almost a third of the way to the tenure review at my new job.
In regards to how to shorten this time, perhaps reviewers and editors could demand fewer revisions. This same question has also been raised in the physical sciences (http://www.nature.com/news/2011/110427/full/472391a.html).
I would enjoy seeing SIAM take a lead and study the impact of these demands on papers.
David F. Gleich
Sandia National Laboratories
Purdue University (as of August 2011)