3 ms·
One course I've been going through lately is the MIT course on Abstract Interpretation ( http://web.mit.edu/16.399/www/ http://web.mit.edu/16.399/www/ ). Progra
by Darmani 15y ago
One course I've been going through lately is the MIT course on Abstract Interpretation ( http://web.mit.edu/16.399/www/ http://web.mit.edu/16.399/www/ ). Programmers tend to break execution into "cases" to reason about it; e.g.: when writing a routine to reverse a string, you might think about the cases where the string is of even or odd length. Abstract Interpretation lets you capture this kind of reasoning precisely, and thereby automate it.
- Xurinos 15y agoThank you for this reference. I was curious if there was some formal work on this subject. This was one of the holes in my studies.
- silentbicycle 15y agoOlin Shivers's "Control-Flow Analysis of Higher-Order Languages" (http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.9.3971 http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.9.39...) is all about using abstract interpretation for compiler analysis. It's focused on Scheme, but is applicable to higher-order languages in general. It's also noteworthy as part of Yale's T / Scheme current (http://paulgraham.com/thist.html http://paulgraham.com/thist.html), which produced a lot of other great stuff: "Orbit" (http://repository.readscheme.org/ftp/papers/orbit-thesis.pdf http://repository.readscheme.org/ftp/papers/orbit-thesis.pdf), "A Tractable Scheme Implementation" (http://repository.readscheme.org/ftp/papers/vlisp-lasc/scheme48.pdf http://repository.readscheme.org/ftp/papers/vlisp-lasc/schem...) and Scheme48 (http://s48.org http://s48.org). For starters!