Monday, September 19, 2005

Time out for reading

Time out for reading

There is a new Half Price books in Berkeley, along Shattuck Avenue a couple blocks west of campus. Curiously, they're in the same space as the dollar store where I got some pans and utensils just after I moved to Berkeley. As part of their opening celebration, they have an additional 20% off. So I wandered in, saw some familiar faces, and picked up a few books:

  • Adventures of a Mathematician (S. Ulam) --

    This autobiographical book is the only one that I've started. There's a preface that describes Ulam's work, which covers a broader range of pure and applied mathematics than I'd realized. And I thought, excellent, this will be very interesting.

    Then I went to the Au Coquelet cafe to read the prologue and sip something warm, and I realized that this will be very interesting. This book didn't start life as an autobiography; initially, Ulam thought to write a biography of von Neumann. But (on page 5):

    When I started to organize my thoughts, I realized that up to that time -- it was about 1966, I think -- there existed few descriptions of the unusual climate in which the birth of the atomic age took place. Official histories do not give the real motivations or go into the inner feelings, doubts, convictions, determinations, and hopes of the individuals who for over two years lived under unusual conditions. A set of flat pictures, they give at best only the essential facts.

    Thinking of all this in the little plane from Albuquerque to Los Alamos, I remembered how Jules Verne and H. G. Wells had influenced me in my childhood in books I read in Polish translation. Even in my boyish dreams, I did not imagine that some day I would take part in equally fantastic undertakings.

    The result of all these reflections was that instead of writing a life of von Neumann, I have undertaken to describe my personal history, as well as what I know of a number of other scientists who also became involved in the great technological achievements of this age.

    What follows is a remarkable book, part autobiography, part biography, and part introspection on the workings of memory and of the mathematical mind.

  • The Education of Henry Adams (H. Adams) --

    I'm not sure where I first heard about this book. I think it was from reading Jacques Barzun, either Teacher in America or A Stroll with William James; that it would be from Barzun is a pretty good guess, though, since Barzun's books are usually crammed with references to books that I decide I want to read (and then quickly lose in my list). Either way, I know in advance to expect more than basic autobiography from this one.

  • Fool's Errand (R. Hobb) --

    A little lighter reading. I enjoyed the first trilogy, but every time I looked for the start to the second trilogy on the local bookshelves, I could only find the second book and on. So now I have something for the next time I feel like a novel.

Right now I'm switching back and forth between Ulam's book and the most recent SIAM Review for my evening reading. This issue of SIAM Review is a really good one, both for technical content and for style. J.P. Boyd, who has written a book on spectral methods which I've mentioned before, has an article on Hyperasymptotics and the Linear Boundary Layer Problem which is both informative and highly entertaining. To quote again:

Unfortunately, asymptotics is usually taught very badly when taught at all. When a student asks, What does one do when x is larger than the radius of convergence of the power series?, the response is a scowl and a muttered asymptotic series!, followed by a hasty scribbling of the inverse power series for a Bessel function. But of course, that's all built-in to MATLAB, so one never has to use it any more.

Humbug! ... Arithmurgy [number-crunching] hasn't replaced asymptotics; rather, number-crunching and asymptotic series are complementary and mutually enriching.

The article refers several times to a longer article (98 pages) entitled The Devil's invention: Asymptotics, superasymptotics, and hyperasymptotics, which I think I would have to put on my reading list just for the title, even if I didn't know that I found the author so entertaining. But given the state of my reading list right now, perhaps it will have to go onto the longer list rather than the shorter one.

Most of my reading time recently has gone to technical material (excepting Tools for Teaching by B. Davis, which is the textbook for the teaching course). This is unfortunate, because what I read tends to strongly influence what things I think about for casual conversation. So if I'm only reading about number theory, Gershgorin disks, and the error analysis of non-self-adjoint PDE eigenvalue problems, I tend to end up talking about those things to anyone who will listen. Those topics are really cool, and I do have buddies who are willing to go have a coffee, or a cider and some pizza, and have a conversation that wanders in and out of technical realms. Nevertheless: a little light reading in the evenings seems to make conversations in the day go more smoothly.