Papers Every Programmer Should Read: None

I like academic papers precisely because of the lack of language-specific details muddying my interpretation of what's going on. I was reading a paper recently describing a SAT-like search problem and the algorithm was described in graph theoretic terms. Very simple, powerful explanation of a very practical algorithm. Implementation was much easier with a graph and rigorous justification than subjectively beautiful source code.

Pure trolling for traffic. Reading source code is a very ineffective way to learn beyond the "micro" level.

I consider erudition, a sense of history, and curiosity to be some of the most important qualities in any professional. Someone who hasn't read classics such as Why FP Matters, the original Lisp papers, the YACC paper, Codd's papers, The UNIX Time-Sharing System etc would show a complete lack of all of those.

That's a lot of hard-won knowledge about difficult problems and new techniques you'll never rediscover on your own.

There are cases where hours' worth of banging your head on a specific implementation won't give you the insight you could glean from minutes' worth of reading a good paper.

- Understanding the practical characteristics of a data structure.

- Understanding the practical characteristics of an algorithm.

- Understanding the principles and motivation behind a design pattern.

- Understanding the principles and motivation behind an architectural paradigm.

Someone who thinks they can learn all they need to know about such topics by looking at source code isn't going to hit their career's apogee much past the "code monkey" mark.

It's a bold statement, sure, but I have to admit that I am not able to refute it.

So, is there, according to you, at least one paper that EVERY programmer should read? Please avoid papers about the difference between "every" and "most" ;)

If no such paper exists, what does it tell about programming? That it is a poorly understood discipline? That would be my guess.

oh look. a post with bold assertions but no evidence. no useful content; just opinion. welcome to hn.

dupe