I'm a little bit confused by Definition 1.64 on page 32 in this book.

This definition says: A prevariety is called quasi-projective variety if it is isomorphic to an open subvariety of a projective variety.

What is an open subvariety of a projective variety? Is it an open subset of a projective variety which endowed with the induced topology and the induced space with functions is again a prevariety? Directly after the definition the book says something different: It says quasi-projective varieties are of the form $(Y,\mathcal{O}_Y)$ where $Y$ is a locally closed subspace of the projective space $\mathbb{P}^n$. What does "locally closed" mean? And what does "subspace" mean? Is that somehow related to linear subspaces of $k^{n+1}$? Then the book talks about some independence on the choice of $X$ which would not be difficult to show. It would be cool if you also could explain how $X$ comes into play.


Edit: The mentioned $\mathcal{O}_Y$ is of the form $\mathcal{O}_{X\mid Y}$ for a closed subvariety $X$ of $\mathbb{P}^n$.

  • 2
    $\begingroup$ Locally closed means closed in a open subset, which is the same as open in a closed subset, which is the same as the intersection of a closed set and an open set. Subspace means simply subspac of the topological space $P^n$. $\endgroup$ – Mariano Suárez-Álvarez Jun 14 '14 at 21:16
  • $\begingroup$ It would be best if you explained what $X$ is. The chances of getting a useful answer and of your question being useful for other people increase immensely if your question is self-contained. $\endgroup$ – Mariano Suárez-Álvarez Jun 14 '14 at 21:18
  • $\begingroup$ I linked the book for reference. But you're right, it's better if it is contained in the question. Therefore, I added it. Btw, thanks for your explanation of locally closed and subspace. $\endgroup$ – principal-ideal-domain Jun 14 '14 at 21:28
  • $\begingroup$ Links to the book may stop working in the future, leaving your question incomprehensible. Always, always ask complete questions :-) $\endgroup$ – Mariano Suárez-Álvarez Jun 14 '14 at 21:34

In principle, the $O_Y$ attached to $Y$ depends on the closed set $X$: indeed, given a locally closed subset $Y$, we know that there is a closed subset $X$ in $P^n$ such that $Y$ is open in $X$, and we define $O_Y$ as the restriction of $Y$ of $O_X$.

Now, there are in general many closed subsets of $P^n$ of which $Y$ is an open subset, and the whole point of the remark is that $O_Y$ does not depend on the choice of $X$: if $X'$ is another closed subset of $P^n$ such that $Y$ is an open subset of $X'$, the restriction $O_{X'}|_Y$ is isomorphic to $O_X|_Y$ as sheaves of rings on $Y$.

You should try to prove this yourself, even if it takes a long time. It is a great exercise in handling the definitions of everything.


Okay, here are some facts.

  1. A projective variety $X$ is a closed subspace of $\mathbb{P}^n$. The canonical open cover of $\mathbb{P}^n$ by $n+1$ copies of $\mathbb{A}^n$ restricts to give a cover of $X$ by the open affines $U_i$ (where $U_i$ is the intersection of $X$ with the $i$th copy of $\mathbb{A}^n$ inside $\mathbb{P}^n$). The sheaf of functions on $X$ is then just the gluing of the sheaf of functions on each $U_i$.
  2. A quasi-projective variety $U$ is an open subvariety of a projective variety $X$. Since $X$ is any closed subset of $\mathbb{P}^n$, $U$ is any open subset of a closed subset (in the subspace topology), which is the same as the intersection of an open subset with a closed subset in $\mathbb{P}^n$. Its sheaf of functions is precisely the sheaf of functions on $X$, restricted to the open subset $U$ (since sheaves are defined on open subsets, there is no ambiguity on how to define this. Here, I mean that on any open subset $V$ of $U$, it's also open in $Y$, so $\mathcal{O}_U(V) := \mathcal{O}_Y(V)$. On the other hand, if you try to restrict a sheaf to a closed subset, there are in general many ways to do this). Hence, by "subspace" they just mean a subspace in the topological sense, ie a subset with the induced subspace topology.
  • 1
    $\begingroup$ There is an ambiguity, and that it does not matter is precisely the point the book is making. $\endgroup$ – Mariano Suárez-Álvarez Jun 14 '14 at 21:32
  • $\begingroup$ @MarianoSuárez-Alvarez He probably meant when $X$ is fixed there is no ambiguity anymore how to define $\mathcal{O}_Y(U)$. $\endgroup$ – principal-ideal-domain Jul 30 '14 at 11:51

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.