Pinned post

I've been seeing a flood of new names, refugees from Twitter. I've been a refugee from Livejournal and Google+, so I know what that's like.

I'm a CS professor, working in algorithms, graphs, and geometry. I also blog and edit Wikipedia. My posts here include links on mathematics, computing, and academic politics, and to my blog, Wikipedia articles, papers, and occasional photos. If you post mathy stuff or I know you elsewhere, I'm likely to follow back; you're also welcome to lurk.


Good Shurik Grothendieck:

James Propp writes (another) account of Grothendieck, as a way of explaining of how a scene from Good Will Hunting on the nature of genius may not be as inaccurate as he once thought it was.

Irritated by repeated claims that spiral galaxies are logarithmic spirals, sourced to pop science instead of expert research, I found "Pitch angle variations in spiral galaxies" (Savchenko & Reshetnikov, MNRAS 2013, which tells a more complicated story. In log spirals, pitch angle is constant, in Archimedean it decreases with radius, in hyperbolic it increases. All three are used as models. But real galaxies have more varied pitch angles that do not match these models.

0xDE boosted

But first, in no small part due to the recent political news, I decided to publish a draft chapter on my blog

It's about randomized response, but specifically how the technique was first applied to measure abortion rates pre-Roe v Wade.

Show thread

Polyomino loops:

Matthew Yuan wonders, if you play billiards in a polyomino and try shooting a ball diagonally from the midpoint of each of the polyomino edges, how to count the number of loops that these billiard paths will link up into. It's somewhat related to the African lusona drawings that I discussed at

0xDE boosted

Today, May 12th, was chosen as a day to globally celebrate women in mathematics . It was the birthday of Maryam (1977-2017), the first and only woman to win the .

The mystery of who wrote a mathematics paper in a special issue, heavily based on the work of the highly-cited special issue editor:

It's by someone who doesn't seem to exist at a college that doesn't exist. But the special issue editor can show emails from the author, so that's something.

Searching MathSciNet for similarly-titled works produces confidence-inspiring journals like _Chaos, Solitons, and Fractals_ and _Fuzzy Sets and Systems_, but this one isn't indexed.

Pandemic at the conference:

ACM CHI, the annual conference in computer–human interaction, was held in a hybrid format with 1900 physical attendees in New Orleans a week ago. It became a coronavirus superspreader event, despite its vaccine and mask mandates.

Terry Tao tries to make mathematical sense of notations like ± or O(...) that specify something partially rather than exactly:

It's a long post, but I think much less technical than most of Tao's posts.

The information Elsevier tracks and resells about the scientists who access its journals:

An EU GDPR personal information request reveals not just dates and times from journal paper reading, writing, and reviewing data but also user names, phone numbers, and bank account information, whether you read the emails from them, and a huge list of the spam newsletters that they subscribe you to.


Digital marbling:

A recent physics-based simulation project for paper marbling, by Amanda Ghassaei, who was also responsible for an origami simulator ( that I linked with a different url a few years ago.

0xDE boosted

Very grateful to be invited to write this book review for the Notices! It was a fun opportunity to explore the wide range of logic puzzles and games that exist in the world, and how they intersect with mathematics. Check it out at

0xDE boosted

Oskar van Deventer has a new mechanical puzzle that requires 4^55 moves to solve:

So, he's got a system that produces a puzzle whose solution length grows exponentially in the number of parts.

Question: Is it possible to come up with a system whose solution lengths are super-exponential in the number of parts?

Stanford mathematics professor Brian Conrad takes a deeper look at the proposed revisions to the California Math Framework (high school mathematics curriculum), finding an "abundance of false or misleading citations", "misrepresentations of facts and evidence", "guidance lacking details essential for implementation or seeming to be opinions rather than evidence-based", and "bias toward data science ... based on misinformation and hype":, via

0xDE boosted

For a \(n\times n\) grid, the lights out game can be considered as solving a linear system over \(n^2\) variables in mod \(2\). It takes \(O(n^6)\) time. Surprisingly, there is a \(O(n^3)\) time algorithm.

Cop-win graph (, now a Wikipedia Good Article. It was the last of my current nominations, so it may be a while until the next. These are graphs on which a cop can catch a robber in a game where they alternate moving on edges or staying put, as I discussed in The GA reviewer made me take out as unsourced a statement that a recognition algorithm of Spinrad 2004 uses time O(mn/log n), faster than the paper's O(n^3/log n). But I still think it's true.

A tutorial on how to use Inkscape to make nice mathematical diagrams quickly enough for real-time note-taking in mathematical lectures:, via

It's from 2019, so doesn't take advantage of newer features of Inkscape released since then. I use Illustrator, but without a site license it's expensive; the other free program many of our students use is Ipe,, more oriented to PDF than SVG but with good TeX integration.

Computer search proves that you can always complete at least one line in standard-size Tetris:

This is true even if the piece order is specifically chosen to make it difficult for you rather than randomly, and even if you are only allowed to rotate and drop pieces without slides or spins after they drop. Via

Mathematics not trademarkable:

Somehow I missed this, but late last year the Supreme Court of Hungary ruled on a lawsuit asking to trademark the Gömböc (, a 3D convex shape that, with only a single stable position, rights itself when placed on a flat surface. (Like a Weeble, but unweighted.) The answer was no: as "an answer given to a question of mathematics" with a shape "necessary for the Gömböc to self-right", it could not be trademarked.

Show older

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!