This is a wish-list of ideas/results that we want to formalize.
This is a hard problem.
Completeness says that everything that is true in ℚ can be proved from
DLO_axioms
. Said differently, if something is true for ℚ then it is
true for every model for DLO_axioms
(because they all have the same
theory).
- we have a definition of
DLO_axioms
, but not ofDLO_theory
inexamples.lean
. - we have a definition of completeness.
- we need to state the completeness of DLO theory as a theorem.
🚀 Next TODO: Define DLO_theory
in examples.lean
.
After that, we can use Vaught’s theorem to prove completeness.
We are close to finishing the proof (assuming Lownheim-Skolem as an axiom).
🚀 Next TODO: Finish the proof of Vaught’s theorem.
This is a map from ℕ to the long strings encoding prime factorization.
This was becoming too hard so we abandoned it. All the partial
definitions we made are currently in godel_encoding.lean
.
There is currently an open issue regarding this here: issues/41
🚀 Next TODO: Close issue #41.
Examples/notes/thoughts: - x<2 in ℝ defines (-∞, 2) - x=y in ℝ defines a line at 45 degrees. - Non-definable: (ℤ, +). ∃x, x+x=x defines {0}. Cannot define {1}. - Is ℤ definable? - Are even numbers (ℤ, +) ∃ y, x=y+y → (ℤ, +) is not o-minimal.
🚀 Next TODO: Not sure. Need to figure out what the Next TODO is.
This uses something called ultraproducts
Los’ theorem implies compactness and Loewenheim-Skolem.
🚀 Next TODO: Define ultraproducts.
Follows from Vaught’s theorem. We would definition of a filter.