Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
Borceux (Handbook of Categorical Algebra, Volume 1, Definition 2.11.1) defines final functors as those functors for which restriction preserves limits, which is the exact opposite of what the nLab says (i.e., restriction along final functors preserves colimits).
Is this just Borceux’s idiosyncrasy? His books generally appear to use conventional terminology. Shouldn’t this be mentioned in the article?
The new version is confusing: “final” appears as a name for two opposite notions!
Maybe we should offer some guide to the terminology. I already mentioned Borceux’s book terminology: cofinal for colimits, final for limits. Mac Lane’s terminology: final for colimits, initial for limits. What other canonical sources should be considered?
I think Borceaux is the odd one out. We can mention his terminology, but we shouldn’t suggest that its use be continued.
I think in the definition of final functor, there is a typo: “ is final if for every object the comma category is……” should be: is final if for every object the comma category is……
Jishen Du
is correct; doesn’t type check. In particular, is the comma category from the functor that picks out to the functor .
added a brief mentioning (here) of the discussion of finality with respect to weighted limits by:
This would be useful to have a citable reference for.
This would be useful to have a citable reference for.
Oh, I see now that this can essentially be cited from Kelly §4.5. Will make an edit..
added mentioning of the characterization of final functors between groupoids (here)
and added pointer to:
In the characterization (proposition 3.1), item 4, shouldn’t the target of be instead of ? Since there is no proof or reference given I can’t be sure, but I see no reason to believe that we can reconstruct the connectedness condition by limits instead of colimits in .
this is taken from Kashiwara & Schapira 2006, Prop. 2.5.2
(which was still more visibly the reference given when this was added in revision 8 in Jan 2010)
I have made it clearer now
1 to 20 of 20