Zulip Chat Archive
Stream: mathlib4
Topic: Ordering on Hahn series
Violeta Hernández (Dec 10 2025 at 04:44):
#25168 added a LinearOrder instance on docs#HahnSeries. Might I ask why it's defined on the Lex type and not directly? I don't believe there is any other reasonable ordering one might give Hahn series. The colex ordering doesn't work, since the value group is only well-founded in one direction. The pointwise ordering does work, but isn't mathematically interesting. @Weiyi Wang @Eric Wieser
Violeta Hernández (Dec 10 2025 at 04:44):
FTR, the lexicographic ordering is the one that gives the isomorphism between surreals and real Hahn series over themselves.
Eric Wieser (Dec 10 2025 at 05:59):
This was discussed in the PR here
Eric Wieser (Dec 10 2025 at 06:00):
This was partly just me saying "deciding whether to make this the order is bikeshedding, let's not let the rest to go waste in the meantime"
Violeta Hernández (Dec 10 2025 at 07:32):
Hm, I think I see the reasoning. Putting global orders on other power series structures would be odd, so we don't do it here either
Eric Wieser (Dec 10 2025 at 08:06):
Odd enough that it made sense to wait for someone to complain that Lex was inconvenient
Violeta Hernández (Dec 10 2025 at 08:15):
Well... I guess I'll have to actually try it out and see how much I can hide it.
Last updated: Dec 20 2025 at 21:32 UTC