I finished the last of the sixteen Tournament of Books contestants (well, except that I couldn’t actually finish The Stranger’s Child). I haven’t commented on the last four, but I read and enjoyed The Last Brother, Salvage the Bones, The Cat’s Table, and The Tiger’s Wife.
Of the four, I enjoyed The Tiger’s Wife and The Cat’s Table the most. Both require some patience, and I didn’t get into them to the extent that I was thinking about them when I wasn’t reading them, but they are worth the effort. The Tiger’s Wife easily beats The Stranger’s Child in the first round, as does The Cat’s Table over Swamplandia! I enjoyed Swamplandia! but it feels like it has been years since I read it, and the story didn’t stick with me like a great book does.
The dog in the photo is our oldest, Nika, who turned fifteen last September. She is having trouble with her hind legs, and often has no appetite, but when we go for walks on the Creek or trails, she’s still as excited and animated as she was when she was a puppy. I’m listening to the A’s vs. Cubs game now, but I think I’ll take her out for a little walk later. The A’s introduced Cuban sensation Yoenis Cespedes earlier today, but he isn’t in the starting lineup. I will be very interested to see how he handles major league pitching, but that probably won’t happen for a few days.
The Morning News Tournament of Books starts next week, and I’m about a third of the way through the only unread volume in this year’s contest, Téa Obreht’s The Tiger’s Wife. In preparation for my posts on the tournament, I wanted to generate a tournament bracket, filled with my choices. I could have fired up Inkscape or my favorite old drawing program, xfig, but drawing something that has such an obvious pattern built into it would be much easier using a programming language rather than moving a mouse pointer around over and over again.
I debated refreshing my Metapost skills, the tool I used to generate my best baseball scorecards, and even wrote a simple Pic (a language written in 1982 by Brian Kernighan, of K&R and awk fame) macro to do it. But I wanted something that would work on the web, and for figures on the Internet, SVG is really the best format (well, unless you’re stuck on Internet Explorer…). Metapost and Pic produce PostScript and PDF, and I wasn’t happy with the way the available PS to SVG image converters mangled the Pic PostScript.
I settled on Python (of course!) and the svgwrite library. Generating an SVG file programatically is pretty easy with svgwrite. You start the drawing with:
svg = svgwrite.Drawing(output_filename, size = ("700px", "650px"))
and then draw stuff onto the page with commands like those found in my draw_bracket function. The all-caps variables are global parameters set at the top of the code to control the size of various elements across the entire figure. I also created a Point class for handling x / y coordinates in the diagram. That’s what start is defined as below (and how I can access x and y via start.x and start.y.
def draw_bracket(svg, start, width, height, names): """ Draw a bracket from (start.x, start.y) right width, down height, to (start.x, start.y + height), placing the names above the horizonal lines. """ lines = svg.add(svg.g(stroke_width = 2, stroke = "red")) lines.add(svg.line((start.x, start.y), (start.x + width, start.y))) lines.add(svg.line((start.x + width, start.y), (start.x + width, start.y + height))) lines.add(svg.line((start.x + width, start.y + height), (start.x, start.y + height))) texts = svg.add(svg.g(font_size = TEXT_SIZE)) texts.add(svg.text(names, (start.x + TEXT_RIGHT, start.y - TEXT_UP))) texts.add(svg.text(names, (start.x + TEXT_RIGHT, start.y + height - TEXT_UP)))
With this function, all that’s left is to design the data structures that hold the data in each column of the diagram, and write the loops to draw the brackets. Here’s the first loop:
# ROUND 1 current_start = Point(MARGIN, MARGIN) for bracket in first_brackets: print(current_start) draw_bracket(svg, current_start, INIT_BRACKET_WIDTH, FIRST_BRACKET_HEIGHT, bracket) current_start = current_start + \ Point(0, FIRST_BRACKET_HEIGHT + FIRST_BRACKET_SKIP)
first_brackets is a tuple of paired tuples, so each bracket above contains the two book titles that should appear on each leg of that bracket.
first_brackets = ( ("Sense of an Ending", "The Devil All the Time"), ("Lightning Rods", "Salvage the Bones"), ... )
The full code can be downloaded at make_bracket.py. The result:
I’ll comment more on my picks after I’ve finished The Tiger’s Wife, and as the tournament starts next week. There were a lot of great books in the tournament, and I wouldn’t be disappointed if some of the later bracket winners in my diagram wound up winning. For me, the most interesting first round bracket is Wil Weaton’s tough choice between Ann Pachett’s State of Wonder and The Sisters Brothers by Patrick deWitt. I chose Sisters, but it was a tough choice, and even though it doesn’t make it past round one on my diagram, I’d be happy to see Wonder win it all.
One other note on the screenshot from my iPad at the top of the post. There are several excellent, non-Tournament books pictured. I highly recommend The Last Werewolf, by Glen Duncan, Neal Stephenson’s Reamde, 11/22/63 (Stephen King), and Eleanor Henderson’s Ten Thousand Saints. These four are easily better than the worst of this year’s Tournament (Hollinghurst, DeWitt and Zambreno).