daniel gratzer (@dannygratzer) header
daniel gratzer (@dannygratzer) profile picture

phd student @ aarhus university. (modal) type theory, (higher) category theory. he/him. 🏳️‍🌈.

Aarhus, Denmark.
Joined July 2013
What is the next step in your career? Come join the department as a PhD student. The next application deadline is August 1. @BirkedalLars and Peyman Afshani have open positions! But you can also use the general call ➡ phd.nat.au.dk/for-applicants…
1
3
921
Huh, so today it clicked for me that the Bousfield-Kan formula for colimits should be understood as the higher-categorical generalization of the classical construction building a colimit out of a (reflexive) coequalizer & coproducts.
1
1
11
1,536
(and indeed, together with some results on 1-cofinality, it is literally a generalization)
1
464
While writing this "synthetic Iris" stuff, one thing I found remarkable: Iris uses Cameras for a sort of 'step-indexed partial commutative monoid'. The definition has a lot of random warts though around how partiality is handled. Some of this is pragmatic
1
4
963
Turns out, this is a bad idea! You can do everything you expect _except_ give a unital completion of a camera. This would be fine, except the main construction one does with cameras generalizes unital completion.
1
1
333
What makes this interesting: it's _provably_ impossible in a constructive setting. Having unital completion without the assumption of decidable support for core implies a weak form of the law of excluded middle.
1
301
Thesis update: drafts for all chapters except preliminary materials and the conclusion exist! Modulo a few random subsections, things are in good shape. Two months left to polish & fill in the remaining text.
2
400
"Cubical type theory can't hurt you" Meanwhile, cubical type theory:
1
7
2
78
3,657
Creating some new copy-pasta with @carloangiuli
3
1
11
1,087
The face of a dog witnessing a horrible crime (me eating a carrot and not sharing)
3
537
Currently drafting an explanation of how to build up Iris -internally- to a modal type theory. Essentially hoping this might produce the definition of Iris @jonmsterling has long asked me for *or* an entry point to modal type theory the Iris folks here at AU have also discussed.
2
1
12
1,141
This thesis is rapidly become A Thing... Starting the draft of the final mainmatter chapter. After that just the intro, conclusion, and bits and bobs I didn't feel like writing the first go through.
3
395
Functional Programming in Lean is now done! This is an online book on using Lean 4 as a programming language that doesn't assume you already know Haskell, Rust, Coq, Idris, or Agda. Thank you to everyone who provided feedback along the way! leanprover.github.io/functio…
7
73
2
345
34,941
5 years in denmark and I remain 100% baffled by Eurovision.
4
11
2,054
tfw you've coauthored a paper too many with @lambdabetaeta
1
1
3
1,692
Congratulations to assistant professor @sophiay135 from @AarhusCrypto, who is one of three recipients of the L’Oréal-UNESCO for women in science 2023 award by @VSelskab 🎉 Read more at cs.au.dk/news-events/news/sh…
1
4
1
38
3,594
daniel gratzer retweeted
A new section, on dual and conjugate fibrations, was added, see kerodon.net/tag/04C0. Also lots of other changes were implemented, especially in kerodon.net/tag/03V7 and kerodon.net/tag/03JB
6
13
2,815
Slides from today's talk on unfolding in implementations of type theory: jozefg.github.io/papers/2023…
7
22
3,269