Friday, December 08, 2006

Not all Nonsense

The talk at IHEP next week has multiplied to talks at Tsinghua and the ITP in quick succession. Should be fun.

In the mean time the weekend promises some rather interesting possibilities, mixing Chinese movie stars and karaoke in one big out-of-tone-glam ensemble.

For now I am bound to link to two stories which have become prominent in the last couple of days.

One of these is from Cosmic Variance, where Joe Polchinski has enlarged on his American Scientist article, answering some of the criticisms Peter Woit and Lee Smolin's books leveled at string theory and the modern theoretical physics community.

This is, unsurprisingly, one of the most level-headed reactions to these books and I look forward to seeing whether the debate can continue with this level of maturity. Unfortunately these threads frequently descend into name calling and the physics is lost somewhere after the first few comments.

Secondly, on a not so impressive note are the claims of a Reading computer scientist to have solved a millenium old problem of dividing by zero. As Clifford Johnson mentions, this sounds like it's straight out of The Onion. Unfortunately this seems to be all too real. The commentators on the BBC site seem to realise, for the most part, that this invention is about as vacuous as the denominator. Mark-Chu-Carrol has written about the best account I've found debunking this bizarre idea. What's most worrying as he says is that there's a class load of kids who not only believe that their teacher is a genius but that the solution he's found to this non-problem is anything other than meaningless.

OK, stars to mingle with, ears to maim.

3 comments:

Cyrus Cygnet said...

Nullity? Nothing to worry about.

Sitting Novation said...

Why is it a millenium old problem? I just did 5 divded by 0 on my calculator and it worked fine and said the answer was zero. If my cheap casio can do it then I'm sure everyone else can...

Jonathan Shock said...

SN,

Clearly it's a millenium old problem because they didn't have casios back then.

The whole thing is particularly bizarre because not only is his solution not useful but having invented his own solution, which is perfectly reasonable, one must throw out the rules of basic algebra.

It's a valid question to ask what happens when you change the basic tenets (axioms) of mathematics to see what sort of a mathematical system you get.

This man's main point is that he has given computers a way to deal with this but, as you've just noticed computers do have a way to deal with it. Your casio has a particularly strange way to deal with it I'm afraid but the point is any computer programmer who has ever encountered an error should write a line that deals with a division by zero. Most computer systems provide the answer as NaN (not a number) and the programmer must then define what the program will do with the error. By calling this object nullity and letting the computer program perform mathematical manipulations on it is even worse because then you have not signalled that there is a problem when there clearly is one.

Looking at the comments on the BBC website it seems that this guy is getting a pretty hard time. He should know better and should certainly be given a firm talking to. However, the main offender here is the journalist who should know better. Anyone writing about mathematics for the BBC should have a little more sense than this and perhaps, if he or she is unsure, should speak to another mathematician or computer scientist.