Suppose is a short exact sequence of quasicoherent sheaves. If two out of three of are locally free, what can we say about the third?
If and are locally free, then need not be. This is a fundamental example.
If and are locally free, then is too. This is important and not too hard.
If and are locally free and finite rank, then is too. This is also useful.
But if and are locally free, and infinite rank, then need not be locally free. The purpose of this post is to give Daniel Litt‘s explanation to me of a simple example of this. I thought this example was not right to include in the Rising Sea (it would be distracting), but I wanted to preserve it for posterity (and for myself).
I will describe a surjective map of free modules whose kernel is not locally free. Viewed geometrically, this provides a surjective map of trivial (infinite-dimensional) vector bundles whose kernel is not a vector bundle. Of course, the kernel is projective.
Let be a projective module over a ring which is not locally free. (Such things exist, see for example the Stacks Project tag 05WG. To be explicit, there and is the ideal (where there are 1’s).)
Then is a direct summand of a free module, say . Writing
gives as a direct summand of a free module with free complement. Then there is a short exact sequence
exhibiting as the kernel of a surjective map of free modules, where the map is the projection.
August 18, 2022 at 1:00 pm
Was the link to the Stacks Project meant to be to a particular tag?
August 18, 2022 at 3:33 pm
Ah yes, thanks, I forgot! 05WG (and 05WH). I’ll fix that now. (Side question that I will ask in my next post: is it outrageous to use “domain” instead of “integral domain” everywhere? I’d mention it at the start of the notes, but of course few people will read that and remember it.) Given the phrases “principal ideal domain”, “euclidean domain”, and “unique factorization domain”, this doesn’t seem unreasonable — but probably it will just confuse some people (?) in return for just saving a few syllables.
August 19, 2022 at 4:40 am
I’m not sure. I think that, if I already remembered that the object you were talking about was a ring, I would get it easily, but if I had momentarily forgotten which side of the algebra/geometry correspondence I was on, I’d start thinking about things like the domain of definition of a rational function. I also think that some people use “domain” for a not-necessarily-commutative ring in which ab=0 implies a=0 or b=0.
My usual preference is to take shortcuts in speech but prioritize being unambiguous in writing.
August 19, 2022 at 8:16 am
I find that all very convincing. I *think* I could be pretty unambiguous, but that’s not quite sufficient. Thanks David!