qmail in particular, and also (if you like math) primegen.
BTW, his code is not easy to understand (quite the opposite) -- there are few source comments, and most of these are snarky, cryptic or both; and not only does he explain almost nothing on a global scale but oftentimes what he is doing might look, at first blush, kinda wrong, obtuse even.
But that's precisely the point -- there are a lot of idioms in there that most C programmers just wouldn't think of, especially if you're into minimalism (when it comes to things like dealing with strings, error handling, forking, etc).
Léon Bottou's implementation of Stochastic Gradient SVMs is the most concise implementation of an SVM that you'll find: http://leon.bottou.org/projects/sgd
"Support vector machines (SVMs) are a set of related supervised learning methods... The standard SVM takes a set of input data, and predicts, for each given input, which of two possible classes the input is a member of... Intuitively, an SVM model is a representation of the examples as points in space, mapped so that the examples of the separate categories are divided by a clear gap that is as wide as possible. New examples are then mapped into that same space and predicted to belong to a category based on which side of the gap they fall on." [wikipedia]
The sqlite test suite is a fantastic example of how to produce solid software. I wish it were used as a case study for all software engineering courses.
I have to disagree. The language spec and the docs on the Lua website are well written but the VM itself is a mess. It's a trainwreck of ambiguously named variables, minimal documentation, and overuse of macros to force inline code.
The Lua VM is a good example of something that isn't well written from a reading / understanding point of view, but is something that is well written from the 'it works' point of view.
This is very Ruby specific (so mostly not "famous" stuff) but if that's your scene, I asked the same question to Rubyists recently and got some good suggestions: http://www.rubyflow.com/items/1817
The Lift framework core has some rather lithe Scala in it. Do mind that it's an actively developed, actively deployed framework - functionality sometimes takes priority ... ;)
BTW, his code is not easy to understand (quite the opposite) -- there are few source comments, and most of these are snarky, cryptic or both; and not only does he explain almost nothing on a global scale but oftentimes what he is doing might look, at first blush, kinda wrong, obtuse even.
But that's precisely the point -- there are a lot of idioms in there that most C programmers just wouldn't think of, especially if you're into minimalism (when it comes to things like dealing with strings, error handling, forking, etc).