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

Patricia Fara - Four Way Interview

Patricia Fara lectures in the history of science at Cambridge University, where she is a Fellow of Clare College. She was the President of the British Society for the History of Science (2016-18) and her prize-winning book, Science: A Four Thousand Year History (OUP, 2009), has been translated into nine languages. An experienced public lecturer, Patricia Fara appears regularly in TV documentaries and radio programmes. She also contributes articles and reviews to many popular magazines and journals, including History Today, BBC History, New Scientist, Nature and the Times Literary SupplementHer new book is Erasmus Darwin.

Why history of science?
I read physics at university, but half-way through the course I realised that had been a big mistake. Although I relished the intellectual challenge, I was bored by the long hours spent lining up recalcitrant instruments in dusty laboratories. Why was nobody encouraging us to think about the big questions – What is gravity? Does quantum mechani…

The Idea of the Brain: Matthew Cobb *****

Matthew Cobb is one of those people that you can’t help but admire but also secretly hate just a little bit for being so awesome. He is professor for zoology at the University of Manchester with a sizable teaching load that he apparently masters with consummate skill. He’s a scientific researcher, who researches the sense of smell of fruit fly maggots; I kid you not!  He’s also an attentive and loving family father but he still finds time and energy to write brilliant history of science books, three to date. His first, The Egg and Sperm Race, describes the search for the secret of human reproduction in the seventeenth and eighteenth centuries and is one of my favourite history of science books, on the period. His second, Life’s Greatest Secret is a monster, both in scope and detail, description of the hunt to decipher the structure and function of DNA that along the way demolishes a whole boatload of modern history of science myths. The most recent, and the subject of this review, is

The Big Ideas in Science - Jon Evans ***

The starting point of a review like this has to be to congratulate the author on his achievement, Jon Evans, because getting all of science into one relatively short book is a massive (and thankless) task. Although inevitably the result is a fairly hectic dash through the material, with limited space for subtleness, Evans manages to make the experience readable and has a light touch that is effective without becoming too simplistic.

There is only one reason this book doesn't get four stars - it's not the quality of the writing but rather the selection of the contents. Of course, there is bound to be plenty of stuff missed out - how else could you get all of science into 269 pages? But the balance is strangely skewed. Chemistry is pretty much omitted, though aspects of chemistry occur under other headings. But for me, the real problem is that physics is really under-represented. It's interesting to use Jim Al-Khalili's recent excellent physics summary title The World Acc…