Recurse Center

Introducing “Paper of the Week”

David Albert

Update 9/22/14: Hacker Schooler Marcus Malka suggested “How to Read a Paper” by Srinivasan Keshav as a companion to Paper of the Week. We’ve included it here for posterity. You can read more about it in PotW #2.

Today we’re introducing the first recurring series on the Hacker School blog. It’s called Paper of the Week. Every Monday we’ll link to a computer science paper that’s worth reading, submitted by a member of the Hacker School community.1

Our goal is to present academic computer science in a practical and accessible way. Reading the right papers can be enriching and enlightening, but there’s a lot of CS research happening and it’s hard to figure out what papers have useful insights into everyday programming. Lots of Hacker Schoolers read and talk about papers, and we thought it would be fun to share some of them with the rest of the world.

We’ll be featuring papers that are relevant to programmers of all levels. Some of them will be Ph.D dissertations, others will be short essays published in journals or just posted to the Web.

If you’re new to academic writing, the language can be dense and intimidating. This is true even for experienced programmers. Don’t despair! Sometimes you will need to read a sentence one word at a time, looking up the definition of each term you don’t understand. Often, it will take multiple reads to grok what the paper is about. Even if you don’t end up fully understanding a paper, the experience of puzzling through it is often worthwhile.

This week’s paper is Growing a Language by Guy Steele. It is a version of a talk he gave at OOPSLA ‘98. At the time, Steele was working at Sun Microsystems on the Java programming language. He was already well-known for his work developing Scheme with Gerald Sussman.

The paper was submitted by Hacker School resident Mel Chua who had this to say about it:

I ran into this during my first programming job as a fresh college grad. I didn’t have a formal CS education, so I didn’t know much about CS theory, and… to be entirely honest, at some level, I didn’t even want to look at it (because what if I looked and found it was Too Hard and I was Not Smart Enough To Be A Programmer?) So I irrationally avoided formal CS theory due to Impostor Syndrome.

However, I loved humor – which was how this paper snuck on my radar under the guise of “check out this great joke!” It wasn’t until partway through that I realized I was laughing at CS theory. Steele’s talk uses natural language to illustrate how programming languages are built up from primitives, but he does so in such a gentle, conversational way that it takes a little while to get the joke of what he’s trying to do. At first, it just sounds like he’s talking funny about topics that aren’t relevant to programming, but stay with it. It is absolutely worthwhile.

Happy reading!

  1. If this gets popular, we may also publish submissions from the public. No promises that we’ll ever do this, but if you feel the urge to send in your favorite paper, you can email