If you are fronted web developer then you should defenitely read jQuery sources. There is a lot of patterns that you could borrow and reuse in your own JS libraries.<p>A good place to start is this interactive code viewer: <a href="http://www.keyframesandcode.com/resources/javascript/deconstructed/jquery/" rel="nofollow">http://www.keyframesandcode.com/resources/javascript/deconst...</a>
There is also a great presentation by Paul Irish ("10 Things I Learned from the jQuery Source"): <a href="http://www.youtube.com/watch?v=i_qE1iAmjFg" rel="nofollow">http://www.youtube.com/watch?v=i_qE1iAmjFg</a>
I'd say the original lisp paper: <a href="http://www-formal.stanford.edu/jmc/recursive/recursive.html" rel="nofollow">http://www-formal.stanford.edu/jmc/recursive/recursive.html</a> by John McCarthy
SQLite probably has some of the cleanest and most elegant C you will ever see: <a href="http://www.sqlite.org/sqlite-src-3070602.zip" rel="nofollow">http://www.sqlite.org/sqlite-src-3070602.zip</a>
There is no single code or project that I could name. Not because there are no open source codes or projects that weren't worth reading, because there are thousands of them. I couldn't name any, because people are different, and what one finds good and worthy code, the other finds rubbish - thus, there can be no single project that would make every reader happy.<p>On the other hand, if you look at the question in a different way, you could say that the code (be it open source or not) every programmer must read before his end, is his own. As one looks at his life in one's deathbed, so should a programmer look at his code.
The Sudoku solver by Peter Norvig <a href="http://norvig.com/sudoku.html" rel="nofollow">http://norvig.com/sudoku.html</a> and not only for the code but also for the excellent essay.
Literate programming is the style of programming that's <i>intended</i> to be read: <a href="http://www.literateprogramming.com/" rel="nofollow">http://www.literateprogramming.com/</a><p>Quite the contrast with the Bourne Shell:
"Nobody really knows what the Bourne shell's grammar is. Even examination of the source code is little help."
– Tom Duff
<a href="http://en.wikipedia.org/wiki/Bourne_shell#Quotes" rel="nofollow">http://en.wikipedia.org/wiki/Bourne_shell#Quotes</a>
If you have only few minutes before execution, and want to fix an old X programming mystery, I would suggest dwm.c: <a href="http://hg.suckless.org/dwm/file/e901e70f69e8/dwm.c" rel="nofollow">http://hg.suckless.org/dwm/file/e901e70f69e8/dwm.c</a><p>It is so easy to understand inner workings of window manager in X reading this code, and you do not need much time for it (>2000 SLOC).
The Linux kernel. If you look at it really closely it isn't so pretty, but its massive size and overall complexity for me (as an undergraduate) was like looking into the grand canyon. What's more impressive is the speed at which things change and that despite breaking all of the traditional rules for "good" software design practices...it works amazingly well.
WordPress - because it's extremely crappy but at the same time insanely accessible. Lots of lessons to learn.<p>Passenger - code looks good even though it's C++<p>Rails 3. Beautifully structured application foundation that's not just a pretty piece of code, it's tested and works.<p>JScheme. Reading this clean implementation and re-implementing it yourself gives a good basis of understanding for lisp.
Based on a blog post that made the rounds a while ago -- <a href="http://tomayko.com/writings/unicorn-is-unix" rel="nofollow">http://tomayko.com/writings/unicorn-is-unix</a> -- Unicorn might be a good read for Ruby-related stuff.<p><a href="https://github.com/defunkt/unicorn" rel="nofollow">https://github.com/defunkt/unicorn</a>
Plan 9 source code (<a href="http://plan9.bell-labs.com/plan9/" rel="nofollow">http://plan9.bell-labs.com/plan9/</a>), mostly for its historical value - created by the same team that made C and Unix - and for an insight on how those people envisioned the evolution of Unix.
I think a very good example of when OOP is genuinely useful and desirable is the Twisted Networking Engine: <a href="http://twistedmatrix.com/trac/browser/trunk/twisted" rel="nofollow">http://twistedmatrix.com/trac/browser/trunk/twisted</a>
Finch is really a clean compiler implementation in C++, it's kind of weird to me when reading its code express nasty things in a clear way in <i>C++</i><p><a href="http://finch.stuffwithstuff.com/" rel="nofollow">http://finch.stuffwithstuff.com/</a>
Probably the most funny source code I encountered was that of the linux kernel, just look for comments.<p>If you just want to die by reading code, you can try to understand how Xen works ;)
(disclaimer, it is actually very elegant)
One of my university professors some years ago said the source of Rogue was one of the C program he read. I've never been able to find it, and I am curious since then. Does anyone?
How about the Lion's book?<p><a href="http://en.wikipedia.org/wiki/Lions_Commentary_on_UNIX_6th_Edition,_with_Source_Code" rel="nofollow">http://en.wikipedia.org/wiki/Lions_Commentary_on_UNIX_6th_Ed...</a><p>Get the source code from <a href="http://minnie.tuhs.org/cgi-bin/utree.pl" rel="nofollow">http://minnie.tuhs.org/cgi-bin/utree.pl</a>
When I downloaded mongrel2 I wanted to take a quick peak at superpoll beyond what was in the blog post about it, an hour later I was still reading the source. It is some of the best-written C I've seen.