Presenting abstract interpretation as a topic in compilation is somewhat akin to presenting dataflow analysis---what is really being given is a set of high-level techniques for analysis. I've thus focused on good overview papers, in the hopes of giving an idea of what is out there and how this stuff works. Note that All abstract interpretation is grounded in semantics---as a result, papers on the subject tend to be extremely dense and contain a lot of notation. I encourage you to focus on the important high-level details of what's going on; it's otherwise all too easy to get lost.

Note that only one of these papers is available on the Web. I've given copies of the others to Saman, and photocopies should be available from him.

The most readable introduction to the subject I have found is the one by

Hughes: Hughes, R. J. M. Compile-Time Analysis of Functional Programs. Research Topics in Functional Programming, David Turner, ed., Addison-Wesley 1990, pp. 117-154.

The seminal paper on the subject, still very readable and dealing with a first-order imperative language rather than a higher-order functional language, is:

P. Cousot and R. Cousot. Abstract interpretation: A unified lattice model for static analysis of programs by construction of approximations of fixpoints. PoPL77, pp. 238-252.

The semantic basis used to derive an abstract interpretation can vary greatly; many papers assume you start with a denotational semantics, for example, even though few real languages actually possess them. An overview of some of the current techniques can be found in:

P. Cousot and R. Cousot. Abstract interpretation frameworks. Journal of Logic and Computation, 4:2, Aug '92, pp. 511--547. http://www.ens.fr/~cousot/publications.www/CousotCousot-JLC-n2--3-p103--179-1992.ps.gz

Finally, this paper was recommended to me by Jamey Hicks as a good example of an application of abstract interpretation.

Alain Deutsch. On determining lifetime and aliasing of dynamically allocated data in higher-order functional specifications. POPL90.