What I read in 2017

2/19 P. G. Wodehouse and C. H. Bovill, A Man of Means

2/13 Philipp K. Janert, Data Analysis with Open Source Tools

Data analysis is a broad topic, and this book is indeed a broad overview. It covers many of the techniques one might need in data analysis, including various graphical techniques (the author emphasizes that when you starting to do exploratory work with a dataset you should do simple things to get some understanding before you pull out high-powered tools, and one of the first things to do when you start looking at a dataset is to look at it), techniques for time series, basic probability and statistics (both classical and Bayesian), reporting systems for business intelligence applications, and machine learning. The book does indeed cover the sorts of open source tools one would expect, like numpy and R and gnuplot and GSL, but a lot of the book also covers things that you can and should do without a computer: guesstimation, dimensional analysis, analytical approximations, and other ways of getting some confidence that the numbers coming out of your computer mean something. The author is frank about which tools and techniques he has found to be useful and which he hasn't; even if you disagree with some of his assessments (as I do; for that matter, I imagine he has changed his mind on a few things in the 7 years since the book was published), it's still helpful to see them. There are also lots of good pointers to more in depth treatments of all of these fields.

Recommended. The author is an ex-physicist who now works with computers, so he speaks my language, but I think I would have found it a good overview regardless.

2/12 Winston S. Churchill, The World Crisis: The Aftermath

The pageant of victory unrolled itself before the eyes of the British nation. All the Emperors and Kings with whom we had warred had been dethroned, and all their valiant armies were shattered to pieces. The terrible enemy whose might and craft had so long threatened our existence, whose force had destroyed the flower of the British nation, annihilated the Russian Empire and left all our Allies except the United States at the last gasp, lay prostrate at the mercy of the conquerors. The ordeal was over. The peril had been warded off. The slaughter and the sacrifices had not been in vain and were at an end; and the overstrained people in the hour of deliverance gave themselves up for a space to the sensations of triumph. Church and State united in solemn thanksgiving. The whole land made holiday. Triple avenues of captured cannon lined the Mall. Every street was thronged with jubilant men and women. All classes were mingled in universal rejoicing. Feasting, music and illuminations turned the shrouded nights of war into a blazing day. The vast crowds were convulsed with emotions beyond expression; and in Trafalgar Square the joy of the London revellers left enduring marks upon the granite plinth of Nelson's column.
Who shall grudge or mock these overpowering entrancements? Every Allied nation shared them. Every victorious capital or city in the five continents reproduced in its own fashion the scenes and sounds of London. These hours were brief, their memory fleeting; they passed as suddenly as they had begun. Too much blood had been spilt. Too much life-essence had been consumed. The gaps in every home were too wide and empty. The shock of an awakening and the sense of disillusion followed swiftly upon the poor rejoicings with which hundreds of millions saluted the achievement of their hearts' desire.

Churchill is famous today largely because of his role in one of the aftermaths of the Great War, but this book doesn't get that far. The fourth volume of The World Crisis takes us from the end of the War up through 1928, including the technical problems of demobilisation, the Paris Peace Conference, the Russian Revolution and Civil War, Irish independence and civil war, the collapse of the Ottoman Empire and the resurgence of Turkey, the redrawing of the map of the middle east, and the rebirth of Poland. A lot of this book is still memoir, what Churchill saw and did as Minister of Munitions and later as Secretary of State for War and then as Colonial Secretary, but there are also large chunks, including all of the parts about the Paris Peace Conference, dealing with events that Churchill had no direct involvement with.

The book ends on a note of hope, but, even without knowing what happened in the future, it seems like a fragile hope. It's clear that the task before the national leaders is to prevent another war, that it's “comparatively easy to patch up a peace which will last for thirty years” and more difficult to create a longer term peace, and that it's impossible to tell whether they've succeeded in even the easy part. The book, as Churchill writes in the first page or two of the preface, is “unhappily for the most part a chronicle of misfortune and tragedy.” Some of the failures were obvious at the time, and others have become clearer over the century since then.

2/11 Alison Bechdel, Fun Home: A Family Tragicomic

We just saw the musical at the Curran in San Francisco (recommended), so naturally I wanted to reread the book.

2/8 Agatha Christie, And Then There Were None

Originally published (in 1939) under an unfortunate name. I reread it because I recently watched the 2015 miniseries, which was well done and largely faithful to the book.

2/4 Matthew Kirk, Thoughtful Machine Learning with Python

I didn't like it very much, I'm afraid. I was hoping for a practical guide to the tools that have now become industry standard (e.g. numpy and Pandas and scikit-learn), but this book was too unfocused, too light on tutorial and practical guidance, with too much digression and throat-clearing about the benefits of OOP and testing. Worse, too many things in it, including some of the digressions, were only half right. A chapter entitled “Probably Approximately Correct Software” is an exhortation about OOP and TDD and the LSP and refactoring, and has nothing to say about Valliant's PAC learning paradigm. The “curse of dimensionality” is a thing, yes, but explaining it as “all of the data points become chaotic and move away from one another” isn't helpful. It is indeed an old saw that economics is “the dismal science,” but not because “we're trying to approximate rational behaviors.” Despite the fact that this book uses Python the author seemed more comfortable with Ruby, and a couple of times even referred to Python modules or packages as gems.

2/2 Agatha Christie, After the Funeral

A Poirot novel from the early 50s, also published as Funerals are Fatal.

1/25 Winston S. Churchill, The World Crisis: 1916–1918

The third volume of Churchill's WWI history/memoirs, taking us through the end of the War. This one focuses much less on naval matters (although it does have two chapters on Jutland, because how could it not?), since Churchill, as punishment for the Gallipoli disaster, was forced from his position as First Lord of the Admiralty in mid 1915.

This is the book that includes Verdun and the Somme and Passchendaele, and Churchill is just as scathing about the tactics of the British and French generals as any anti-war poet. (In fact one chapter even uses a line from Siegfried Sassoon as an epigraph.) Some of this was self-serving, as Churchill was writing with his reputation in mind, arguing that his earlier ideas were right and that there should have been more effort put into tanks, or forcing the Dardanelles, or exotic amphibian attacks, instead of attempting again and again to send infantry in frontal attacks against fortified positions defended with barbed wire and concrete and machine guns. Even if it can't be taken at face value, though, in retrospect it seems largely right.

The anatomy of the battles of Verdun and the Somme was the same. A battlefield had been selected. Around this battlefield walls were built—double, triple, quadruple—of enormous cannon. Behind these railways were constructed to feed them, and mountains of shells were built up. All this was the work of months. Thus the battlefield was completely encircled by thousands of guns of all sizes, and a wide oval space prepared in their midst. Through this awful arena all the divisions of each army, battered ceaselessly by the enveloping artillery, were made to pass in succession, as if they were the teeth of interlocking cog-wheels grinding each other.

1/11 Miran Lipovača, Learn You a Haskell for Great Good!

An introductory Haskell tutorial. A big part of the book is basically about showing that monads aren't so scary.

1/10 Sir Arthur Conan Doyle, The Case-Book of Sherlock Holmes

“I fear that Mr. Sherlock Holmes may become like one of those popular tenors who, having outlived their time, are still tempted to make repeated farewell bows to their indulgent audiences. This must cease and he must go the way of all flesh, material or imaginary … And so, readers, farwell to Sherlock Holmes! I thank you for your past constancy, and can but hope that some return has been made in the shape of that distraction from the worries of life and stimulating change of thought which can only be found in the fairy kingdom of romance.”

This collection was the third time that Doyle attempted to write a last Sherlock Holmes story. This time he succeeded: The Case-Book of Sherlock Holmes was published in 1927, and Doyle died in 1930.

1/8 George O'Connor, Ares: Bringer of War

And for a graphic novel about the god of war, what other war could it be? This is basically a much abridged Iliad, starting with the wrath of Achilles and ending with the death of Hector, focusing on the role of the gods. The reason for Achilles's refusal to fight is left vague; Briseis isn't mentioned.

1/8 The Iliad (tr. Caroline Alexander)

  • Wrath—sing, goddess, of the ruinous wrath of Peleus' son Achilles,
  • that inflicted woes without number on the Achaeans,
  • hurled forth to Hades many strong souls of warriors
  • and rendered their bodies prey for the dogs,
  • for all birds, and the will of Zeus was accomplished;
  • sing from when they two first stood in conflict—
  • Atreus' son, lord of men, and godlike Achilles.
  • 1/1 Sir Arthur Conan Doyle, His Last Bow

    I imagine the title story in this collection was intended yet again as the last Sherlock Holmes story. It wasn't — there was one last collection in the late 20s — but it is the last story in internal chronology. Holmes, now in his 60s, is pulled out of retirement for one last case. The story was written in 1917 and takes place on 2 August 1914, and Holmes's case is just what you'd think given those dates.

    1/1 George O'Connor, Aphrodite: Goddess of Love

    Sixth book in the Olympians series of graphic novels, about half of which is Aphrodite's origin story (roughly the version from the Theogony). As the cover illustration suggests, this book also includes the wedding of Peleus and Thetis and the Judgment of Paris. Turns out it's a bad idea to have Eris, goddess of discord, as one of your wedding guests, but leaving her off the guest list isn't a great idea either.

    1/1 George O'Connor, Poseidon: Earth Shaker

    Fifth book in the Olympians series of graphic novels. Among other things it includes the the founding of Athens, the cyclops episode from the Odyssey, and Theseus and Ariadne and the Minotaur.

    Matt Austern