Zulip Chat Archive

Stream: mathlib4

Topic: Naming of universal property lemmas


Eric Wieser (Aug 11 2025 at 21:53):

/poll Which of these is called lift_comp_ι?

  • lift (g ∘ ι) = g
  • (lift f) ∘ ι = f

Eric Wieser (Aug 11 2025 at 21:54):

And of course the more important question, what should the other one be called?

Eric Wieser (Aug 11 2025 at 21:55):

We have precedent here, but the precedent is to call the other one ι_comp_lift, which is definitely not a reasonable reading of the naming guide.

Aaron Liu (Aug 11 2025 at 22:05):

can you give some examples?

Aaron Liu (Aug 11 2025 at 22:06):

I'm having trouble inferring the types in lift (g ∘ ι) = g

Aaron Liu (Aug 11 2025 at 22:07):

oh I figured it out

Eric Wieser (Aug 11 2025 at 22:07):

g : BigHom (F X) Y, f : LittleHom X Y, ι : LittleHom X (F X)

Eric Wieser (Aug 11 2025 at 22:08):

I guess strictly speaking (lift f) ∘ ι = f is actually (lift f).toLittleHom.comp ι = f, which maybe solves the naming issue

Kenny Lau (Aug 11 2025 at 22:41):

lift_app_comp_ι would be the other one :smile:

Aaron Liu (Aug 11 2025 at 22:46):

Eric Wieser said:

I guess strictly speaking (lift f) ∘ ι = f is actually (lift f).toLittleHom.comp ι = f, which maybe solves the naming issue

Strictly speaking lift (g ∘ ι) is actually lift (g.toLittleHom.comp ι)

Aaron Liu (Aug 11 2025 at 22:48):

but sometimes LittleHom and BigHom are the same type

Michael Stoll (Aug 12 2025 at 08:41):

lift__comp_ι and lift_comp__ι? :grinning_face_with_smiling_eyes:

Eric Wieser (Aug 12 2025 at 08:44):

lift_comp_ι and comp_lift_ι, ignoring the infix-ness?


Last updated: Dec 20 2025 at 21:32 UTC