After just missing advancing out of Google Code Jam Round 1A, I tried again in Round 1B and again in Round 1C, but with no success. I guess I haven’t adapted well to the new interactive problems. In both cases, I got the first problem solved and then got stuck on the second problem, which was interactive in both cases. The interactive problems require your submission exchange messages with another program to work out the solution, which makes it a bit different to debug.

In the interest of advancing, I should have moved on to the third problem or just did the easier subset of the interactive problem. But I’m only doing this for the fun of the challenge, and it was fun to eventually work out the solution, even though it took me a little longer than the allotted time.

I tried the 2.5-hour Round 1A Friday night and just missed the cut-off for advancing. The round started a few minutes late as the problems site was overloaded at first. When it did respond, I got the problems in a different order, with the last problem listed first. Thinking it was the first (and therefore easiest) problem, I started on it without noticing the difficulty scores. That “Alien Rhyme” involved finding pairs of words with common suffixes. There were a few gotchas to the greedy approaches, so I was lucky to work out a correct algorithm, but it took me a few tries to get a decent data structure for organizing the words by common suffixes. I ended up with a vector of maps, one per suffix length with each map itself mapping a suffix to a set of words. Amazingly, it worked.

Next I tried the Pylons problem which boiled down to finding a complete path through a graph. I couldn’t think of a good approach for the large 20×20 constraint, so I went for partial credit with a brute force solution which would work for the small 5×5 constraint. After that was submitted I noticed that the large graphs were dense enough to have many complete paths, so maybe a randomized brute force approach would work. I added a little randomization and resubmitted. It turned out I had the right idea, but I didn’t randomize it enough (only the starting points and not the node visitation order). So I still only got partial credit, but with a penalty for making a second submission.

The remaining problem was an interactive one and I had too little time left for the overhead of testing it locally, so I didn’t attempt it. I could see that the solution may require solving a Chinese Remainder problem, and I did solve it the next day. It’s nice that the site enters practice mode when the competition is over, so we can test out ideas later.

The good news is that I get to compete in Round 1B!

This year’s Google Code Jam started yesterday with the 27-hour Qualifying Round. I managed to get all the problems correct for a score of 100/100 along with about 1000 of the 35,000+ participants. You only needed a score of 30 to advance, so many who could have done better probably stopped when they had enough points to qualify.

The problems always get harder with each round, but this year’s qualifying round seemed easier than usual. The first two problems only took a few minutes each, and I almost stopped there since I knew I already had 30 points.

The third problem, Cryptopangrams, was also easy to figure out, but the solution required doing math on 100 digit integers, which I couldn’t readily do in C++. I briefly looked around for a simple big-integer library to include but decided to relearn enough Python to use that. It felt silly googling things like whether array indices start at 0 or 1 (it’s 0 in Python), how to comment out a line (‘#’ character), and how to do integer division (‘//’ operator). I never did come across a good Python cheat sheet; instead I had to wade through various introductory teaching pages. Fortunately, there was no real time pressure and it was a simple problem.

All the problems have at least one small/easy test case and one large/hard test case. For the fourth problem, I could figure out the easy case quickly but had no idea how to solve the hard case. After re-reading the problem statement I realized the parameters were constrained enough (looking for at most 15 errors in a 1024 bit string) for me to solve. These are most fun problems to solve: when you start with no clue and keep looking at different angles until you find a solution.

The other complication for the fourth problem was that it was interactive. Instead of the usual read-problem-input/write-solution sequence, an interactive problem requires multiple queries and responses before the final solution can be determined, and that makes testing your code much trickier. Google nicely provides a testing Python script, which I eventually got working in my CLion environment. Had to use chmod to make testing_tool.py executable and had to prefix it with #!python for it to run. Then I set up a CLion target for the interactive_runner.py script which would launch the testing tool and my code and get them to talk to each other. Maybe I can remember enough of it to get through the next round when time is more constrained.

The Code Jam results list is a bit harder to navigate than it used to be. I wanted to look at other contestants’ solutions to the Cryptopangrams problems to see if any of them used C++. You can still look at submitted solutions, but when you do so you lose your place in the list, so you have to start over at the top of the list with each try. I looked at a few entries in the top 20 and none used C++ for that problem. Most used Python, some Java and one Ruby.

Just to avoid the embarrassment of going a calendar year without a blog post, here is a summary of some of my 2017 highlights. It’s not that I’ve been silent, but I’ve mostly switched to microblogging on Twitter (@xangregg) for sharing updates.

Tie-dye marathon

Some summers I keep track of a tie-dye “marathon” where I see how many days in a row I can go wearing a different tie-dye each day. This year I made it 32 days. Most of the images are on Twitter. Here are days 13 – 24.

Google Code Jam 2017

I tried both the code jam and the distributed code jam this year. In the regular code jam I made it to the round of 3000 before bombing out with a tied-for-last score. I keep forgetting the later rounds often require ready-made advanced optimization algorithms. I did better in the distributed code jam, advancing to the round of 500, which was enough to win my third code jam T-shirt. The distributed code jam is tricky since you submit code that is run on a distributed system of 100 CPUs. This year it helped that I built a test harness that used 100 threads for a better simulation of the actually process communication issues.

Packed Bars

Trying to find a decent way to show data with many categories in a skewed distribution, I created a new chart type called packed bars. Here’s an example showing costs of billion-dollar disasters before this year.

I presented packed bars in a poster at the IEEE VIS conference in Phoenix, and there are now implementations in JMP, R, Excel and D3.js. Ironically, the JMP script is the weakest implementation pending the arrival of JMP 14 in March 2018. It’s only great such data and has some learning curve, but skewed data is pretty common, especially in quality control (defect counts) and finance. I hope packed bars can become useful to others.