Skip to main content

The Golden Ticket: P, NP, and the Search for the Impossible – Lance Fortnow ***

There is good and bad news early on in this book about the P versus NP problem that haunts computing. The good news is that on the description I expected this to be a dull, heavy going book, and it’s not at all. Lance Fortnow makes what could be a fairly impenetrable and technical maths/computing issue light and accessible.
The bad news is that frustratingly he doesn’t actually tell you what P and NP mean for a long time, just gives rather sideways definitions of the problem along the lines of ‘P refers to the problems we can solve quickly using computers. NP refers to the problems to which we would like to find the best solution’, and also that he makes a couple of major errors early on, which make it difficult to be one hundred percent confident about the rest of the book.
The errors come in a section where he imagines a future where P=NP has been proved. This would mean you could write an algorithm to very efficiently match things and select from data. Fortnow suggests that our lives would be transformed. This is slightly cringe-making as fictional future histories often are, but the real problem is that he tells us that the algorithm would make it possible to do two things that I think just aren’t true.
First he says that from DNA you would be able to identify what a person looks like and their personality. Unfortunately, these are both strongly influenced by epigenetic/environmental issues. Anyone who knows adult identical twins (with the same basic DNA) will know that they can look quite different and certainly have very different personalities. And they will usually have been brought up in the same environment. Fortnow is forgetting one of the oldest essentials of computing – it doesn’t matter how good your algorithm is, GIGO – garbage in; garbage out.
The other, arguably worse error is that he says that it will be possible to have accurate weather forecasts going forward X days. This is so horribly wrong. He should have read my book Dice World. The reason you can’t predict the weather at all beyond about 10 days is nothing to do with the quality of the model/algorithm, it is because the system is chaotic. Firstly we just don’t know, and never can know, the initial conditions to enough decimal places not to deviate from the real world. When Lorenz first discovered chaos it was because he entered the starting values in his model to 4 decimal places rather than the 6 to which the model actually worked. It soon deviated from the previous run. We can’t measure things accurately enough. The other problem is that the weather system is so complex – hence the slightly misleading title of Lorenz’s famous paper Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas? – that we can’t possible take into account enough inputs to ever have so good a model as to go forwards that far. Sorry, Lance, it ain’t going to happen.
For the rest, the first half or so of the book goes along pretty well, gradually opening up the nature of P and NP, the problems that are of interest and the ‘hardest’ NP complete problems. I found the main example, used throughout, a hypothetical world called Frenemy where everyone is either a friend or enemy of everyone else confusing and not particularly useful, but Fortnow gets plenty of good stuff in. After that it’s as if he rather runs out of material and it gets a bit repetitious or has rather tangential chapters.
Overall, despite the flaws, a much better and more readable book than I thought it was going to be – but probably best for maths/computing buffs rather than the general popular science audience.

Hardback 

Kindle 
Using these links earns us commission at no cost to you
Review by Brian Clegg

Comments

Popular posts from this blog

Lucy Jane Santos - Four Way Interview

Lucy Jane Santos is an expert in the history of 20th century leisure, health and beauty, with a particular interest in (some might say obsession with) the cultural history of radioactivity. Writes & talks (a lot) about cocktails and radium. Her debut book Half Lives: The Unlikely History of Radium was published by Icon Books in July 2020.

Why science?

I have always been fascinated by the idea of science especially our daily interactions with and understandings of science – especially in a beauty context. I could spend hours pondering the labels of things on my bathroom shelf. What is 4-t-butylcyclohexanol (as a random example)? Do I really understand what I am putting on my face and spending my money on? Would it change my purchase habits if I did?  

Why this book?

This book came from an accidental discovery – that there was a product called Tho Radia which contained radium and thorium. I found out about it because I actually bought a pot of it – along with a big batch of other produc…

Rewilding: Paul Jepson and Cain Blythe ****

Those who are enthusiastic about saving the environment often have a mixed relationship with science. They might for example, support organic farming or oppose nuclear power, despite organics having no nutritional benefit and requiring far more land to be used to raise the same amount of crops, while nuclear is a green energy source that should be seen as an essential support to renewables. This same confusion can extend to the concept of rewilding, which is one reason that the subtitle of this book uses the word 'radical'.

As Paul Jepson and Cain Blythe make clear, though, radical change is what is required if we are to encourage ecological recovery. To begin with, we need to provide environments for nature that take in the big picture - thinking not just of individual nature reserves but, for example, of corridors that link areas allowing safe species migration. And we also need to move away from an arbitrary approach to restricting to 'native' species, as sometimes…

Is Einstein Still Right? - Clifford Will and Nicolas Yunes ***

If there's one thing that gets a touch tedious in science reporting it's the news headlines that some new observation or experiment 'proves Einstein right' - as if we're still not sure about relativity. At first glance that's what this book does too, but in reality Clifford Will and Nicolas Yunes are celebrating the effectiveness of the general theory of relativity, while being conscious that there may still be situations where, for whatever reason, the general theory is not sufficient.

It's a genuinely interesting book - what Will and Yunes do is take experiments that are probably familiar to the regular popular science reader already and expand on the simplified view of them we are usually given. So, for example, one of the first things they mention is the tower experiments to show the effect of gravitational red shift. I was aware of these experiments, but what we get here goes beyond the basics of the conceptual experiment to deal with the realities of d…