Senior Lecturer at @ImperialEEE researching programming languages and hardware design. Personal account. Also @wicko3.bsky.social

United Kingdom
Joined May 2009
John Wickerson retweeted
ACM TOPLAS has four new paper tracks: - PL Squibs (Guest Editor: Colin Gordon) - PL Experiments (Guest Editor: @wicko3) - Tools, Systems and Practitioner Reports - Surveys and Tutorials Check out my editorial on these tracks in the latest issue! dl.acm.org/doi/10.1145/37850…
1
4
16
1,218
Amazing initiative by @imperialeee! 💯 Had a lot of fun discussing with the students on the future possibilities of building compliant-by-design ML models! Thanks @wicko3 for organising! 🙏🏼
Big thank you to Eleonora Giunchiglia @e_giunchiglia for a fascinating “Meet the Profs” talk at @imperialeee about her research into compliant-by-construction ML models!
3
12
1,985
Big thank you to Eleonora Giunchiglia @e_giunchiglia for a fascinating “Meet the Profs” talk at @imperialeee about her research into compliant-by-construction ML models!
1
1
11
2,587
John Wickerson retweeted
PACMPL (POPL, PLDI, OOPSLA, ICFP) papers should have a limit of 10 pages. We already write them this way (section 2 is basically the paper at this point) and all the details should go in appendix. Writing this way would make our work accessible outside of our community.
9
3
3
65
15,382
A good day for folks who like runs of numbers!
1
284
Delighted to have our first “Meet the Profs” seminar of the term, featuring Ayush Bhandari (@AB2World) who showed us how to extract a useful signal just from the noise of quantisation! Simultaneously mind-blowing and beautifully explained. @imperialeee
1
1
7
489
Congratulations to Philippa Gardner, Professor in @ICComputing, who has received a @bcs Ada Lovelace Medal 🥇 Professor Gardner received the award in recognition of her work on language specification, software verification and bug detection ⬇️ imperial.ac.uk/news/257553/c…
1
8
1
28
3,302
John Wickerson retweeted
Amazing graphical representation of a neural net, never seen anything like it.
47
3,360
230
19,645
1,642,249
John Wickerson retweeted
I’ve done a lot of formal methods “technical sales” - trying to scope projects for Galois clients that solve their problem and fit into a budget. I gave a talk last week on some things that I learned along the way: mikedodds.github.io/files/ta…
13
95
11
385
37,165
Did I just wake up after a 400-year nap?
4
29
2,271
John Wickerson retweeted
We have a new record! The Great Internet Mersenne Prime Search verified that this massive number is indeed prime. Which also means that a new perfect number has been discovered. The last record held for an unusually long time—almost 6 years. We’ll see how long this stands.
77
710
122
6,398
441,593
John Wickerson retweeted
As the author of this PDF, it's been interesting seeing people guess at the rationale behind its design. However, the rationale had nothing to do with theory vs practice, and everything to do with pragmatically coping with an unaccommodated disability in academia. (1/16)
Compilers was was known to be the hardest CS class at Cornell which was hard as it is. We were handed a 8-page PDF at the start of sem for a language spec we'd be implementing by the end of sem, split into 6 parts. On part 5, the median was a 0/100 and most the class failed.
123
1,424
181
15,874
3,400,845
John Wickerson retweeted
Wow, I appreciate @acm_ccs being frank about review ethics! #CCS
7
53
14
178
54,815
John Wickerson retweeted
Compilers was was known to be the hardest CS class at Cornell which was hard as it is. We were handed a 8-page PDF at the start of sem for a language spec we'd be implementing by the end of sem, split into 6 parts. On part 5, the median was a 0/100 and most the class failed.
638
686
479
11,784
5,539,496
Machine-assisted mathematics is super-interesting. Here, Marijn J.H. Heule was looking for S(5) and used * 14 CPU years to solve the SAT problem, * 2 million gigabytes of proof demonstrating S(5)=160, and * 32 CPU years to check the proof. arxiv.org/pdf/1711.08076
3
7
41
3,545
An interesting paper from @emeryberger et al., showing that, in contrast to prior work, (in my words) energy use across programming languages is a proxy for how long a program takes to execute, and that other factors don't meaningfully affect energy usage. arxiv.org/abs/2410.05460
4
16
1
55
5,216
John Wickerson retweeted
I have fully funded #PhD studentships (including international studentships) available at @EdinburghUni in the areas of Computer Architecture or Compilers. Details are here: jobs.ac.uk/job/DKB486
2
1
3
1,176
Occasionally in mathematics a statement that just has to be true turns out to be false. A paper appeared on arXiv today that disproves a well-known conjecture in probability called the bunkbed conjecture. 🧵 arxiv.org/abs/2410.02545
17
149
28
1,017
125,366