I had the pleasure of reading a draft of a fascinating Part II project dissertation on computational aspects of descriptive complexity theory this week, written by a very talented student at my college. I won’t say much about it, of course, but it was one of those projects that goes far beyond the material we teach in the course—although I have opinions about the limitations of the Computer Science Tripos as a tool for actually teaching people computer science, I must say that it does a very good job getting out of the way from students who need space to run and learn.

By the way, I read the thesis on my white plastic iMac and compiled my feedback using an ancient version of OmniOutliner. Converting the outline to a PDF, moving it to my modern laptop by creating a local network, and sending it by email is a simple enough workflow. OmniOutliner can even read and write OMPL, which I can then feed into Bike if I like—or convert to any format I want using XSLT. That is the beauty of a standard: interop spanning decades.

Last week, I wrote about a pleasant surprise concerning the results of my LICS 2025 paper: one of two annoying side conditions from the main theorem can be dropped, and I conjectured that I could find a way to remove the other side condition. I spent much of the work working on the latter, and unfortunately I have come up a bit empty. There is some real difficulty with the higher coherences here, and although I have tried to attack them from several directions, they are not budging.

I am thinking that I will just make the one improvement prior to the camera-ready submission, and leave the other one for a future paper should I ever solve it.

I finished Adrian Tchaikovsky’s science fiction debut Children of Time last week—my initial impressions were a little hesitant, but I have to say that the quality of the writing improved rapidly as I went onward. There is something a little strange about Tchaikovsky’s style that takes you out of the fictional world and makes you overly aware of reality: he is constantly make analogies that are sensible to you and me but would not make sense to the characters in the book. I don’t think this is necessary even when you are writing about spiders, but I suppose I am learning a bit about my own taste. In any case, I loved the book. I am completely terrified of spiders, but I can recommend this to any arachnophobe—I will write more in my actual review later.

I started reading the sequel, Children of Ruin, and straightaway I can say that the writing quality has improved at least by a factor of two. With Ruin, I find Tchaikovsky hitting a strong rhythm, and his depiction of the alien world Nod is as enchanting as his characterisation of humans (and their arachnid comrades) is compelling.