[Merged by Bors] - chore: Move Bernoulli's inequality#9456
[Merged by Bors] - chore: Move Bernoulli's inequality#9456YaelDillies wants to merge 2 commits intomasterfrom
Conversation
The first versions for ordered semirings do not need to import so much but: * they have no obvious place to go * nobody needs them that early (in particular, it seems nobody needs them without also needing the ordered field version) Part of #9411
Why can't they just stay where they are? Is it such a bad place? |
|
They could, but all other lemmas in |
|
New files are cheap, IMO |
|
Okay but what do I put in this new file? the three Bernoulli inequalities for rings and I leave the field one in |
I would go for
and add some comments pointing back and forth. |
The first versions for ordered semirings do not need to import so much but: * they have no obvious place to go * nobody needs them that early (in particular, it seems nobody needs them without also needing the ordered field version) Part of #9411
|
Pull request successfully merged into master. Build succeeded: |
`Algebra.GroupPower.Lemmas` used to be a big bag of lemmas that made it there on the criterion that they needed "more imports". This was completely untrue, as all lemmas could be moved to earlier files in PRs: - #9440 - #9442 - #9443 - #9455 - #9456 - #9457 - #9459 - #9461 - #9463 - #9466 - #9501 - #9502 - #9503 - #9505 - #9551 - #9553 - #9720 - #9739 - #9740 - #9805 - #9806 - and this one There are several reasons for this: * Necessary lemmas have been moved to earlier files since lemmas were dumped in `Algebra.GroupPower.Lemmas` * In the Lean 3 → Lean 4 transition, Std acquired basic `Int` and `Nat` lemmas which let us shortcircuit the part of the algebraic order hierarchy on which the corresponding general lemmas rest * Some proofs were overpowered * Some earlier files were tangled and I have untangled them This PR finishes the job by moving the last few lemmas out of `Algebra.GroupPower.Lemmas`, which is therefore deleted.
`Algebra.GroupPower.Lemmas` used to be a big bag of lemmas that made it there on the criterion that they needed "more imports". This was completely untrue, as all lemmas could be moved to earlier files in PRs: - #9440 - #9442 - #9443 - #9455 - #9456 - #9457 - #9459 - #9461 - #9463 - #9466 - #9501 - #9502 - #9503 - #9505 - #9551 - #9553 - #9720 - #9739 - #9740 - #9805 - #9806 - and this one There are several reasons for this: * Necessary lemmas have been moved to earlier files since lemmas were dumped in `Algebra.GroupPower.Lemmas` * In the Lean 3 → Lean 4 transition, Std acquired basic `Int` and `Nat` lemmas which let us shortcircuit the part of the algebraic order hierarchy on which the corresponding general lemmas rest * Some proofs were overpowered * Some earlier files were tangled and I have untangled them This PR finishes the job by moving the last few lemmas out of `Algebra.GroupPower.Lemmas`, which is therefore deleted.
The first versions for ordered semirings do not need to import so much but:
Part of #9411