jsmorph (@jsmorph) header
jsmorph (@jsmorph) profile picture
Austin
Joined November 2008
The GRH work by @__alpoge__ timeline represented by Fable.
1
1
3
325
"It is natural to feel that this is an end, in my case a loss of a principal guiding light, but what I saw during the course of progress so amazed me that I believe it is one dawn across science of something far greater." --@__alpoge__
1
1
542
"It is natural to feel that this is an end, in my case a loss of a principal guiding light, but what I saw during the course of progress so amazed me that I believe it is one dawn across science of something far greater." --@__alpoge__
1
6
864
Replying to @guillaumeclaret
In that paper, neither Gallina → Scheme nor Scheme → bytecode are verified. The Rust VM is unverified. No machine-checked correctness theorem covers final artifacts. LeanExe uses verified compilation and kernel-checked theorems connecting supported artifacts to their source computations. For bare-metal images, those theorems extend to execution from boot or reset in explicit machine models. This page summarizes some artifact-level proofs: github.com/jsmorph/leanexe/b…
12
Full IEEE 754 for floating point (no Real)
1
21
Replying to @krismicinski
This "leanexe" dialect is a small subset designed to make executable verification feasible while still allowing for the full power of Lean (and mathlib, etc) in the proofs. Example: github.com/jsmorph/leanexe/t…
4
272
Replying to @TaliaRinger
One goal of leanexe is to have a (small) verified runtime. Here's an example of this Lean dialect compiled to a verified microcontroller image. The executables are verified. Fun. github.com/jsmorph/leanexe/t…
1
176
Replying to @satnam6502
Why not? Could you offer some sort of ZK evidence if you feel like you need to keep the proofs proprietary?
1
31
Replying to @satnam6502
Where are the Lean proofs?
1
68
Replying to @ElucidationsPod
LeanExe (subset of Lean with a verified compiler to WASM and other targets)? github.com/jsmorph/leanexe/t…
1
2
262
I had almost typed bugs/incompleteness!
2
29
yah, specs can (and often do) have bugs
1
1
99