Linkage

Minhyong Kim’s physicsinspired approach to Diophantine equations (G+). New from Quanta.

Comparing singleblind vs doubleblind conference reviews (via). An experiment where halves of the WSDM 2017 program committee were assigned to different processes. Well known authors, and authors from wellknown institutions, got both more nonblind reviewers wanting to read their papers and higher nonblind review scores. Also, oddly, the blinded reviewers requested to read more papers overall.

Covering a hexagon by equilateral triangles. Six triangles of the same side length as the hexagon suffice. But if you have seven, can you use smaller triangles? An unsolved problem from Soifer and Karabash in 2005.

Wikipedia should restore its support for MathJax, according to this nowclosed Christmas wishlist poll entry.

Excerpt from a new biography of pioneering US codebreaker Elizebeth Friedman (via).

How short can a tour that visits all the things on a regular polyhedron be?

MathSciNet is now a cloud atlas to the mathematics literature (G+). Meaning that instead of having mirror domain names the service will be replicated across the web under a single domain. Bonus links: MathSciNet’s tool for finding matches to text references and a commandline tool for searching several other bibliographic databases (via).

Open access to the Bulletin of the Institute of Combinatorics and its Applications.

Clean up and prettyprint svg files. Together with some judicious handediting this can lead to significantly smaller and cleaner vector graphics files, if you care about that.

Dice become ordered when stirred, not shaken (via). “Vigorously twisting” a cup full of dice can quickly arrange them into a neat packing, whereas tapping or shaking the cup takes much longer.

Melbourne Street Art. Shot with my cellphone because I neglected to bring the camera this time.

Don Knuth’s Christmas Lecture 2017 (G+, via). Knuth talks about counting partitions of rectangles into smaller rectangles.

Sorting algorithms revisualized (via). Pretty but also informative.

In case you needed any more evidence that trying to use MathML in place of LaTeX only causes trouble, check out the horrible BibTeX generated by Elsevier for one of their papers (G+, more G+ comments). A single \(\chi\) has somehow become 13 lines of nonfunctional mathml, with the wrong kind of chi encoded as TeX inside it, and with the mathml angle brackets also TeXencoded for extra ugliness. And look out for pinnipeds!

Covering the unit sphere by neighborhoods of great circles requires the total (geodesic) widths of the neighborhoods to be at least \(\pi\) (G+, via). This 3d variant of Tarski’s plank problem was open for 44 years. See also the arXiv and published version of the original paper by Jiang and Polyanskii, as well as another higherdimensional generalization of Tarski’s problem by Keith Ball (and thanks to the commenters for the extra links).