A Day In The Life Of A Mathematics Undergraduate

Via Mathematics Weblog, here is an unusual promotional film for the 4-year MMath degree course at the University of Warwick:

The Evolution of the Bicycle

Theoretical physicist Freeman Dyson has said a good many perceptive things in his time. Here is one of his sayings that I had not come across before:

You can't possibly get a good technology going without an enormous number of failures. It's a universal rule. If you look at bicycles, there were thousands of weird models built and tried before they found the one that really worked. You could never design a bicycle theoretically. Even now, after we've been building them for 100 years, it's very difficult to understand just why a bicycle works - it's even difficult to formulate it as a mathematical problem. But just by trial and error, we found out how to do it, and the error was essential.

I suppose one of the complicating factors for bicycles is that they have to work in close conjunction with human bodies, with all their foibles and peculiarities. Anyhow, next time you see a bicycle, try to imagine it as the survivor of a process in which thousands of other "weird models" were weeded out.  Come to think of it, the cyclist too has been subjected to the same process but in their case many millions of weird models had to be weeded out (though, to judge from the cyclists I have seen, some of the weird models still survive to this day).

From Today The Keyboard Is Dead

In this Google Tech Talk,  David J.C. MacKay of Cambridge University presents Dasher, a piece of software that appears to make the computer keyboard obsolete.   It is like a computer game in which you use the mouse to run through a fractal pattern of squares holding the letters that you want to enter.  The sizes of the squares are relative to the frequency of the letters given the last few characters entered.  Keypresses and mouse clicks are kept to an absolute minimum.   The idea seems to have everything going for it: it requires no new hardware, it has a several obvious applications (MacKay mentions people with disabilities, hand-held computing, mobile phones, and Chinese and Japanese text entry), and the software is free (GPL) and available for Windows, Linux and Macs.  Unfortunately the project website appears to be down at the moment.  I will go over to my Ubuntu system and see if there is a version available under Synaptic that I can install and play with.

Update on Mon 2007-07-09 by Tristram Brelstaff

I have now had a chance to play with using Dasher on my Ubuntu laptop.  Writing with it is a very strange experience:  recursive, fractal, exhilarating.  You get a real sense of the infinite number of other possible texts sliding past as you move the mouse down your particular sequence of characters.  Very Jorge Luis Borges.

I understand that, for most people,  writing with Dasher would be slightly slower than using a keyboard.  However, it would still seem a good idea to familiarise yourself with using Dasher because it, or something like it, could well become the norm for data entry in keyboardless systems.  And besides, a few more years of development could well make it faster than a keyboard.

One question comes to mind: how useful would Dasher be for programming languages? 

The Dasher Project site is now back up, so Iwill be able to download and try out the Windows version (as soon as my daughter comes off our Windows PC).

Evidence-Based Management Fads

Over the years, in the course of my work I have attended many management courses. Consequently I was interested to read this by Derek Lowe:

...   The contempt that most of the scientific staff has for "modern management techniques" is hard to underestimate. Problem is, we're used to having to prove our hypotheses, and show data (with appropriate controls, yet) in support of them. But I've suspected for years that most of the management fads that sweep through the world have nothing to back them up at all, ...

Lowe is talking about management fads in the drug industry but much the same holds for the software development industry. 

I particularly remember being shown a graph supposedly proving the effectiveness of Fagan inspections but which consisted of just a wiggly rising line.  The actual data points that the line was drawn through were not shown, so it was impossible to get any idea of the quantity or scatter of the points, or of the statistical significance of the rise in the line.  It might be that the person who drew the graph was ignorant of what is needed to convince people trained in science, or it may be that they were deliberately trying to  cover up the inadequacy of their data.  Either way, the presentation was seriously weakened by that graph.  This a pity, as I now understand that Michael Fagan gathered quite a lot of statistical data which he used to convince the management of IBM of the the effectiveness of his technique.  It was just that we weren't shown any of that data.

I suspect that the data that the SEI CMM framework was based on would not stand up to close scrutiny.   At least, I cannot remember seeing any convincing graphs in any of the CMM presentations or books.

What String Theorists do on Saturdays

Over at Asymptopia, Clifford Johnson gives a glimpse into how he spends his weekends.  From last Saturday:

I sat around puzzling over factors of two and minus signs.  Essentially the whole day.  There were a few stops here and there to do a few things, like a quick blog post in the morning over breakfast, think about what I might do during the day, get some laundry started, and so forth, but eventually I got completely pulled back into the matters that had been on my mind all week.

From this Saturday:

Things got worse and worse through the morning, as I realized that many things I so wanted to be right about the next stage of my computation (which perhaps I’ll tell you about one day) were in fact not going to work.  Not even close.