I'm looking for the simplest possible example (one that's easy to remember) for the situation described in the title. More precisely I'm looking for the following example:

A (probably has to be singular) algebraic surface $X$ (2-dimensional, reduced, integral, finite type over $k$ algebraically closed) and a complex of coherent sheaves $\mathcal{F}^\bullet$ on $X$ whose image in $D(X)$ is perfect but not quasi-isomorphic to a strictly perfect complex. (Bonus points for a simple toric surface example).


  1. A strictly perfect complex is an object in the derived category $D(X)$ isomorphic to a bounded complex of vector bundles
  2. A perfect complex is an object in the derived category $M \in D(X)$ satisfying that for every point $x \in X$ there's a open zariski neighborhood $x \in U$ over which $M_U \in D(U)$ is perfect ("locally strictly perfect").
  • 2
    $\begingroup$ If I recall correctly you need a scheme without an ample family of line bundles, and this rules out all quasi-projective examples. $\endgroup$ Jan 7, 2017 at 18:25
  • 3
    $\begingroup$ I am pretty sure that on every quasi-projective scheme over a field, e.g., your curve above, every perfect object is strictly perfect. This should be in SGA 6, but you might first look in Thomason-Trobaugh. $\endgroup$ Jan 7, 2017 at 18:26
  • $\begingroup$ oh shoot. Then let me revise. $\endgroup$ Jan 7, 2017 at 18:44
  • $\begingroup$ Any such example will fail to have the "resolution property". If memory serves, there is an example of a (non-separated) scheme failing the resolution property in the article of Edidin, Hassett, Kresch, and Vistoli. $\endgroup$ Jan 7, 2017 at 21:11

1 Answer 1


Perhaps it is silly to post this as an answer. My answer to your other question also answers this question.
Can the homological dimension of a coherent sheaf explode along a formal deformation? (is the resolution property hereditary for formal deformations?)
That example is smooth, integral, and 2-dimensional over a field. However, it is not separated.

  • $\begingroup$ Thanks for reminding me about this! I was very worried about this issue once. Now that I know that there's a dichotomy (ala the other question) I'm much more relaxed. $\endgroup$ May 17, 2017 at 13:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.