You might think, “One more science book. Ho hum.” But you’d be missing out on a very entertaining book full of bits of knowledge worth knowing. This is written in a light, almost silly tone. In fact, at first I thought it was written with a 10-year-old audience in mind, which irritated me a bit. But as I got into it, I realized it presented some sophisticated ideas and science in very accessible ways for anyone. It consists largely of anecdotes of discovery which can be fascinating, like the woman who discovered she could smell Parkinson’s disease or the border guards who learned too late that they were training dogs to detect plastic wrap instead of the contraband inside the wrap.
This fascinating book about a pioneering surgeon was especially enjoyable because it educated me about things I had never known anything about. In very readable prose the author has mixed World War I history, personal anecdotes about individual servicemen, and medical innovations in plastic surgery in just the right proportions. The extensive notes at the end testify to how thoroughly it was researched. The author holds a doctorate in the history of science and medicine. Yet the book is not the dry academic tome some historians seem to favor. It deals with a grisly subject, severe disfigurement, but maintains a light, upbeat tone, much like Sir Harold Gillies, the main figure of the book. She sticks to the subject matter of facial reconstruction without diverting too much into biography, which I appreciated. If you’re not comfortable looking at facial disfigurement, skip the photos in the midsection, but the text should pose no problems.
Just heard another idiotic news report. “Queen Elizabeth was interned today.” I’m glad to hear she got a job as an intern. Maybe she can get hired on full time if she does a good job. What they meant to say, of course, is that she was interred. A different station reported on an earlier part of the pomp and pageantry when there was an equestrian procession. He said “Here comes the Calvary.” He meant cavalry. These are English announcers butchering their own language. I don’t want to hear any criticism of us Yanks by these limeys about our own butchering, which is at least as bad.
As a dedicated member of the American Cryptogram Association, I like to solve ciphers and I enjoy writing programs for that purpose. Regular readers will have read some of my posts on the subject. Recently I tried an experiment to modify my programs to make them faster. It didn’t work very well.
In its simplest form computer solving can be described as two steps: 1. try a decryption and 2. test the result. This process is repeated, usually until a solution is found or the user stops it. It can be done by running through a word list or by randomized guesswork such as hillclimbing. My idea to speed it up is very simple: don’t test the entire decryption; just test the beginning.
The way my programs usually do the testing is to score every tetragram (4-letter group) for its frequency in English or other target language. The more high-frequency tetragrams in the trial decryption, the more likely it is to be near the solution. I found that if I just test the first four tetragrams instead of the whole thing, it does save time. I found a strict cutoff score that allows 99.3% of valid decryptions through for a full decryption, but cuts off up to 90% of false decryptions depending on the cipher type. I thought this would improve my solving time by 90%. It didn’t turn out that way.
The first problem is that this method doesn’t reduce the number of decryptions, only the number of test cycles. For many cipher types, it is necessary to decrypt the entire thing before you have the beginning to test. It turns out that the decryption phase takes longer than the testing phase. The time savings were there on several types, but so minimal that it wasn’t wroth it. Remember, it’s not 100% accurate, so that means that every now and then it would cut off a valid solution, too, which is a limitation.
The second problem is that this only works for solving methods that rely on a 100% solution. For example if the decryption method is to decrypt using a word list as a key source, you expect that when the correct key word is tried, the solution will be 100% valid English, including the beginning. It is safe to assume that if the beginning has bad tetragrams, one should go on to the next key. But if one is using hillclimbing, that is not so. You might make a change to the key that improves the overall score, and should be preserved, but doesn’t necessarily produce good text at the beginning. My method won’t work at all on a hillclimber, simulated annealing, etc. It also doesn’t work on types with numerals or with Playfair unless your frequency data is modified to deal with the numbers or extra X’s for the same reason.
I tried solving the first problem by extending the method to the decryption process. For many types, such as Vigenere, it is possible to just decrypt the beginning. So I did that, decrypting only the beginning and testing it. There were some savings in time, but surprising to me, it was only around 1% and still had the same risk of cutting out the valid solution. Apparently interrupting the normal loop of decryption to test the start slows down the whole process, especially for a type like Vigenere, that produces quite a few valid-looking beginnings even with wrong keys. The cost of the extra test cycle balanced out the savings from culling out some full decryptions.
Another limitation in the method is that it doesn’t work at all with transposition types. They have the same number of high-frequency letters as normal text, so nearly every trial decryption will pass the test and go on for the full-length decrypt/test cycle. In the end I decided to chuck the whole idea. It still seems like it ought to be possible to streamline the trial decrypt/test process. One way to do it it if you have a crib is to test whether the crib appears in the trial decryption. Similarly, if your decryption process involves reducing the text to a simple substitution, test to see if the pattern of the crib appears.
Bradbury, the author, and this book are both iconic symbols of science fiction. But the book really isn’t science fiction per se. As Bradbury himself says in the introduction, it’s myth. A bugaboo of mine is how booksellers and even libraries lump together science fiction and fantasy. I consider them separate genres. This book would clearly fall into the fantasy half since there is very little science in it. Bradbury makes little effort to portray Mars in ways that are remotely plausible to today’s audience. It is covered with water-filled canals, has sufficient oxygen for people to breathe, and is already populated by happily married Martian couples much resembling humans only with crystal hair and triangular doors. It often sounds silly, and in truth, it mostly is.
But the book isn’t intended to be hard science fiction like, say, The Martian. Bradbury uses Mars as metaphor, recreating the despoliation of the Americas by the Europeans, for how our earthly society could be so much better, or so much worse. Many later sci-fi books have done the same thing, but this book paved the way. It is rather amusing at times, as well as disappointing to some extent, to see how inaccurately the author foresaw the future. Of course it’s easy to see in hindsight, but Bradbury posits the families of the future to look mostly like the families of middle America in the 1950s. Women are happy to stay in the kitchen and go to the beauty parlor. All astronauts are men and most of them smoke cigars. Vehicles on Mars are all gas hogs with fossil fuel rocketed up from Earth. Come on, Ray, electric cars have been around since the 1890s. You could have done better even back then. By today’s standards, the book seems rather juvenile, but it deserves its place in the pantheon of pioneering science fiction.
We’re in the middle of a heat weave here in California. Almost every weather reporter has been talking about triple digits. The idiocy that I can’t stand is when they say – and most of them do – that “it’s going to be over triple digits tomorrow.” No, it’s not. Triple digits is the range 100 to 999. To be over triple digits it would have to be at least 1000 degrees. What they mean is it’s going to be over double digits, or it’s going to be in triple digits. They don’t know the difference between over and in. I think the concept is over their heads when it should be in.
I read Ng’s second book five years ago and found this one to have the same strengths and weaknesses. It begins with a mystery. Lydia, a teenage girl, drowns in a lake one evening, apparently alone. Was it an accident? A murder? A suicide? But for the next 200 or so pages the author leaves that mystery alone and begins an overlong and plodding history of Lydia’s dysfunctional family. The theme is clearly people’s failure to tell others their true feelings and thoughts and how that can cause harm. The mystery is explained at the end and it’s a rather clever twist I didn’t foresee.
Before that, though, the author paints a depressing story of what she seems to be suggesting is racism toward Asians (although she uses the term Orientals in keeping with the custom of the times depicted). I grew up in the fifties and sixties, the time frame in that part of the book, and found it to be an inaccurate portrayal, displaying a rather whiny victim mentality. Just like with the other book, I nearly gave up on it halfway through, and, like the other one, was eventually glad I didn’t. Still, in the end it felt more like a completed school assignment than an enjoyable read.