Zulip Chat Archive

Stream: Is there code for X?

Topic: Norm in quadratic extension


Edison Xie (Jun 27 2024 at 13:31):

I'm currently working with quadratic extension of , namely ℚ(√a), however the method I used to construct it is by Algebra.adjoin ℚ (√a) and working with this is really hard since lean doesn't even know this is a fieldso my first question is how to express the element in ℚ(√a) as x + y *√a? And how do I tell lean that norm in quadratic extension of is precisely x^2 - ay^2?

Riccardo Brasca (Jun 27 2024 at 13:34):

Writing down a definition is a subtle problem. What are you goals?

Riccardo Brasca (Jun 27 2024 at 13:35):

Mathematically Algebra.adjoin ... is of course what you think it is, but I would strongly suggest to use what we call a "characteristic predicate" rather than a concrete definition.

Riccardo Brasca (Jun 27 2024 at 13:35):

By this I mean introduce something like IsQuadratric a on a field K.

Riccardo Brasca (Jun 27 2024 at 13:37):

You can have a look at how cyclotomic fields are set up in mathlib to have an idea.

Edison Xie (Jun 27 2024 at 13:41):

Riccardo Brasca said:

Writing down a definition is a subtle problem. What are you goals?

I've got two goals: 1.express all element in Q(sqrt(a)) in the form of x + y sqrt(a) and 2. show that Algebra.norm in the quadratic extensions are just (x + y sqrt(a))*(x - y sqrt(a))

Edison Xie (Jun 27 2024 at 13:41):

Riccardo Brasca said:

You can have a look at how cyclotomic fields are set up in mathlib to have an idea.

thanks for the tips

Riccardo Brasca (Jun 27 2024 at 13:43):

For a more elementary approach you can follow docs#Zsqrtd

Riccardo Brasca (Jun 27 2024 at 13:45):

BTW I think we really want the theory of quadratic fields in mathlib (it is ridiculous that we have cyclotomic fields and not quadratic fields).

Edison Xie (Jun 27 2024 at 13:51):

Riccardo Brasca said:

BTW I think we really want the theory of quadratic fields in mathlib (it is ridiculous that we have cyclotomic fields and not quadratic fields).

Yeah I know! It's really annoying!

Kevin Buzzard (Jun 27 2024 at 17:16):

Bourbaki develop the theory of quadratic algebras; given a commring R and two elements a,b, we can define R[a,b] to be R[X]/(X^2-(aX+b)). This can be used to make extensions by zero, quadratic fields, orders in them etc etc.


Last updated: May 02 2025 at 03:31 UTC