Because of the heat wave currently keeping Belgium in its grip (preventing me from doing research…) I decided to finally make the last fixes for a new feature for the Stacks project that has been in the make now for a while: detailed tag edits.

Git has a feature called git blame, that tells you when a certain line in a file was last edited and by whom. Starting from this idea Johan has built stacks-history that collects all this information throughout the lifetime of the Stacks project for all tags. If you add a little processing code to get this in the database in a meaningful format and some new website code, and you can check exactly when a tag has been changed!

If you want to see it in action, point your browsers to the history page for the tensor-Hom adjunction.

It’s not perfect (it might misinterpret some changes etc.) so if you spot mistakes, please do tell. And if you have suggestions to make this feature (or the Stacks project in general) better, please do tell!

The late Alexander (Sasha) Rosenberg, has a webpage collecting (some of) his work. Of course, most of his works were available through the preprint server of the Max–Planck Institute, but now things are easy accessible. It only went online a few days or weeks ago I think, and I guess it is maintained by his son Leo. There is also a book containing selected papers, which seems to be the print version of the website I linked.

Enjoy!

A little over two years ago I started my Atlas of $\mathop{\mathrm{Spec}}\mathbb{Z}[x]$. Today I added another map, from a (forever draft?) book of David Mumford and Tadao Oda.

I urge you to check out the website of David Mumford, who, despite having switched from algebraic geometry to vision in the 80s maintains an excellent overview of all his work in geometry. In particular, he recently posted this draft version of a book which is over 40 years old, which contains a TeX’ified version of his map of the affine line over the integers on page 120.

Yes, I am still alive. Today I would like to announce a new series of lectures in our graduate student seminar. We are going to talk about Hodge-to-de Rham degeneration, and the proof of Deligne–Illusie of this. Feel free to join in if you are interested, and somewhere in the neighbourhood of Antwerp.

The last post in the music category dates from almost 3 years ago. And even longer ago I started a not so long-lived series on computational composition using Strasheela. Time to change this situation.

Last week as a programming finger exercise I decided to implement a brute-force first species counterpoint solver in Python, enumerating all legal solutions and then scoring them. The first part is finished now, and can be seen at GitHub.

There are several directions in which I would like to take this little project. First of all it needs to be faster, which might be feasible by multi-threading (something I’d like to try just for getting some experience with multi-threading in Python). Observe that the code certainly hasn’t been written with speed in mind, I went for conceptually clear code not some super-fast C-type implementation. Then of course, I want to implement the scoring system. But before that, there are some rules that aren’t implemented yet (e.g. hidden fifths and octaves come to mind) and some of the rules have an ugly implementation.

Going to second species will be a significantly harder task: one has to interpret the function of a note correctly.

Feel free to chime in with suggestions! I have discovered a similar project and I should definitely try to implement something like melodypy.com but for now you’ll have to look at plaintext and hear the notes in your head.

I just gave a lecture on examples of spectral sequences in algebra and algebraic geometry for the ANAGRAMS seminar and without further ado, here are the notes. Nothing new or insightful in there, just some facts and examples taken together to help me during the lecture.

A long time ago I wrote about how to produce Čech cohomology groups in LaTeX, while in the previous post there is the Čech-to-derived spectral sequence making an appearance, and when I took the first screenshot in that blogpost I realised there was something wrong with the positioning of the exponent. Compare:

to

The first one is the naive approach, obtained by using

\DeclareMathOperator\cHH{\check{\HH}}
\DeclareMathOperator\HH{H}

The problem with it is that the accent raises the height of the box, thereby raising the exponent too. But visually speaking the $q$ and $p$ should be on the same height. This is done by using the \smash macro (which already exists in Plain TeX, and is not part of mathtools which I mistakingly thought up to today) which basically annihilates the height from its argument in all layout computations. It can create a horrendous mess, but here it is the solution.

To make sure that our macro is as generally applicable as possible we do insert the height of the argument which will then act as the total height of the result using the \vphantom macro:

\newcommand\cechit[1]{\smash{\check{#1}}\vphantom{#1}}

\DeclareMathOperator\cHH{\cechit{\HH}}
\DeclareMathOperator\HH{H}


which results in the second screenshot.

The irony is that if I wouldn’t be using macros, there wouldn’t be a problem, as TeX will not treat it as one high box if you hardcode everything. Oh well…