Parsers and parsing accessories. Para-academic. Rootless cosmopolitan redneck. Writes for @Status451Blog. 'Juvenile but not white supremacist' -- @splcenter
Belgium
Joined July 2007
- Tweets42.1K
- Following1.5K
- Followers12.9K
- Likes15.2K
Pinned Tweet
Turns out the seventh Loff-Moreira-Reis conjecture (JCSS 2020) is probably wrong and there's an ordinary parsing expression grammar for even-length palindromes after all.
I have a constructive, AI-assisted proof candidate, not yet independently reviewed. 🧵
The July construction has internal tests, but no independent review or end-to-end Lean or Rocq proof. Please try to break the source-machine, scheduling, or PEG expansion arguments.
Proof, code, and audit notes: github.com/abiggerhammer/peg…
Also added an explainer for those who aren't kept up at night by unanswered questions in the formal languages literature. If you need some background on why this is interesting, see github.com/abiggerhammer/peg…
All right fine I've been summoned.
As usual, the actual problem is tapdancing down the middle of the extremes being debated here. To be able to see it, though, the one thing you must absolutely not do is treat the software landscape as static. It's emphatically not.
By the way, Laurie could have consulted, say, @maradydd or @sergeybratus who wrote the paper she is completely misinterpreting. They are not hard to find. Both are people I know, and I am pretty confident that both of them will agree with me.
But eliminating a bug class from a component is not the same thing as eliminating the process that creates new trust boundaries across components.
Langsec is a ratchet only if you keep hold of the ratchet.