Peter Norvig's “pytudes” for Advent of Code 2020

Always look forward to these so I can compare my solutions afterwards. Sometimes it&#x27;s almost frustrating to see how clean and concise the mess I made could have been, but I do learn a lot. It&#x27;s like he gets as close to golfing as possible, without the obfuscation.<p>Too bad he punted on 20... one of the few difficult problems this year. Usually the hard problems are where you <i>really</i> see the contrast between his solutions and the rest of us plebs. Do I sense a hit of frustration in his comment? ;)<p>&quot;too tedious for too little reward....&quot; &quot;sea monster as characters inelegant&quot;

How do you algorithmically code like this? In my software engineering career, I am a plumber at best. Would fail hard at anyone who would ask me to code anything more than fizz buzz under time pressure. That said, I&#x27;ve built some incredible production worthy and robust systems that move mountains as a plumber. I am damn good at that.

This is very well-written code. I don&#x27;t think anyone would bring into question Norvig&#x27;s technical skills, but I wouldn&#x27;t think someone like him would spend much time at his job writing code, or have much free time to program as a hobby.<p>As someone who aspires to a position similar to Norvig&#x27;s (directing research at an organization that is frequently applied in real life), I think it&#x27;s really cool to see him having time to do stuff like this.

If you like this, you will love <a href="https:&#x2F;&#x2F;github.com&#x2F;mebeim&#x2F;aoc&#x2F;tree&#x2F;master&#x2F;2020" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;mebeim&#x2F;aoc&#x2F;tree&#x2F;master&#x2F;2020</a><p>The guy created very clean python solutions, and very insigthful walkthrough (although a bit wordy some times). It was my go to repo after validating my solutions on AoC (he managed to publish daily during the whole challenge !).

This is another vote of confidence for type annotations. There still seems to be a debate between highly experienced Pythonistas whether typing is good or not.<p>In this case I think they work really well, and aren&#x27;t stuck in places where they&#x27;re not needed.

HTML view of the noteboook, in case Github is failing to load it for anyone else: <a href="http:&#x2F;&#x2F;archive.today&#x2F;Xzyse" rel="nofollow">http:&#x2F;&#x2F;archive.today&#x2F;Xzyse</a> (originally from <a href="https:&#x2F;&#x2F;htmtopdf.herokuapp.com&#x2F;ipynbviewer&#x2F;" rel="nofollow">https:&#x2F;&#x2F;htmtopdf.herokuapp.com&#x2F;ipynbviewer&#x2F;</a> , just didn&#x27;t want to HN hug them).

Probably the most indicative part of Norvig&#x27;s code beauty is the requirements file [0]. With python developers being infamous for importing obscure&#x2F;questionable libraries for even a simple task, I love how Norvig solved all these with just native libraries and the staples numpy&#x2F;plt.<p>[0]: <a href="https:&#x2F;&#x2F;github.com&#x2F;norvig&#x2F;pytudes&#x2F;blob&#x2F;master&#x2F;requirements.txt" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;norvig&#x2F;pytudes&#x2F;blob&#x2F;master&#x2F;requirements.t...</a>

I just find Norvig&#x27;s style of &quot;functional Python&quot; lovely in its own way (with noted disregard of Pep8 and other &quot;best practices&quot;)

It&#x27;s heartwarming to see he skipped d20p2.

I thought I spotted a bug on the very first day, part 2.<p><pre><code> def day1_2(nums): &quot;Find 3 distinct numbers that sum to 2020, and return their product.&quot; return first(x * y * z for x, y in combinations(nums, 2) for z in nums &amp; {2020 - x - y} if x != z != y) </code></pre> The last line (if x != z != y) returns true when x and y are equal and z is different.<p>But x and y are already constrained to be different since nums is a set and itertools combinations picks distinct elements from the iterable. If the test had been x != y != z it would have been a bug.

This was precisely the kind of thing I was looking for. I&#x27;ve done AOC for a few years, and did OK this year (got 2 stars on all but 3 days, and one star on all but 2), but being able to compare my ramshackle, simplistic code to something like this is the project for a month - break down one a day.... I&#x27;ll do this in february, so thanks for posting that, OP!

&gt; cat = &#x27;&#x27;.join()<p>Love it. There&#x27;s so many bits of Python I&#x27;ve written where I should have had something like this.

Day 5 part 1 solution is interesting. I had implemented it the obvious way doing a binary search. I didn&#x27;t realize you could represent that string as binary and get the equivalent integer. That&#x27;s really cool however I am not sure I get the idea behind why that actually works

Can someone please explain how `for y in nums &amp; {2020 - x} ` in his day 1 solution?

The &#x27;timing&#x27; section is disappointing. All very clean though.

Norvig&#x27;s code is gold, even back from the times of PAIP (which is still definitely a must read).<p>His &quot;poker&quot; course (on YouTube) is also golden standard.<p>Why? The background in math, basic data structures and logic, so one know the range and boundaries of what is possible. Plus careful attention to details, like it is in any art.

Even as a software engineer who studied mathematics, I just can&#x27;t get excited about the type of problems found in Advent of Code and other such things, like Project Euler. Those problems amount to little puzzles and tricks and algorithms that have little to no context. What I find interesting about computing and programming is the representation and exploration of ideas.<p>Also, I can&#x27;t help but wonder. Why does the Director of Google Research and a computer scientist use his personal time to code, what looks like exclusively, in Python rather than other interesting languages? Is it a bit of marketing for Google and&#x2F;or his book, or does he just like Python?