Skip to content

add SemiNormVect (originally based on Martin's PR) - #426

Open
cmcq wants to merge 1 commit into
ScriptRaccoon:mainfrom
cmcq:main
Open

cmcq wants to merge 1 commit into
ScriptRaccoon:mainfrom
cmcq:main

Conversation

@cmcq

@cmcq cmcq commented Oct 20, 2026

Copy link
Copy Markdown

This adds SemiNormVect (no NormVect yet). All properties are decided.

@ScriptRaccoon ScriptRaccoon added the data additions and updates to the database label Oct 20, 2026
@ScriptRaccoon

Copy link
Copy Markdown
Owner

Thank you for the PR! I will review it in the next days. But I will already submit a few quick comments in a moment.

Comment on lines +101 to +102
- property: effective congruences
proof: The midpoint argument for <a href="/category/Ban">$\Ban$</a> works here, except that we skip the step of showing $R$ is closed.

@ScriptRaccoon ScriptRaccoon Oct 20, 2026 •

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

When refering to other proofs, use labels and references.

Have a look how this is done e.g. in Mon.yaml with the property "regular subobject classifier" and the file CMon.yaml that references this proof.

The same remark applies to other proofs here that refer to Ban.

Comment on lines +191 to +193
- property: regular quotient object classifier
proof: >-
$\IC$ with its usual norm has exactly two regular quotients, $\{0\}$ and $\IC$. If an object $\Psi$ classified regular quotients, then $\Hom(\Psi, \IC)$ would have cardinality $2$. But this is impossible: if it contains $f \neq 0$, then it contains the distinct elements $0,f/2,f$.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For some proofs you refer to Ban because they are the same, but for some you repeat the argument, even though it is very similar. Why?

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sorry this was referring to the proof of regular subobject classifier

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe Ban should be added as a related category, in both directions.


satisfied_properties:
- property: locally small
proof: The forgetful functor to $\Vect_{\IC}$ is faithful, and $\Vect_{\IC}$ is locally small.

@ScriptRaccoon ScriptRaccoon Oct 20, 2026 •

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Having a faithful functor to a locally small category only settles locally essentially small.

Of course, this property is trivial nonetheless. I would phrase it like:

There is a forgetful functor $\SemiNormVect \to \Vect$ and $\Vect$ is locally small.

name: category of semi-normed vector spaces with linear contractions
notation: $\SemiNormVect$
objects: semi-normed vector spaces over $\IC$
morphisms: linear maps $f$ satisfying $|f(x)| \leq |x|$ for every $x$

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
morphisms: linear maps $f$ satisfying $|f(x)| \leq |x|$ for every $x$
morphisms: linear contractions, i.e. linear maps $f$ satisfying $|f(x)| \leq |x|$ for every $x$

proof: >-
The forgetful functor to vector spaces preserves finite limits and filtered colimits. We just need to show that the semi-norms agree. The case of empty limits is trivial. Now consider filtered diagrams $(X_i)$ and $(Y_i)$ with the same index category.

For equalizers, consider maps $f_i, g_i : X_i \to Y_i$ with equalizers $E_i\subseteq X_i$, and an element $x\in\colim_i E_i$. So $|x|$ is the infimum of $|x_i|_{X_i}$ taken over its representatives $x_i\in X_i$ such that $f_i(x_i)=g_i(x_i)$. Since the diagram is filtered, this is the same as the infimum over all representatives, which gives the semi-norm in $\colim_i X_i$. Hence the semi-norm agrees with that of the equalizer of $\colim_i f_i$ and $\colim_i g_i$.

Copy link
Copy Markdown
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for parallel arrows always use \rightrightarrows

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

data additions and updates to the database

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants