Computer Science Papers Every Developer Should Read

The article says computer science papers help developers understand key concepts in depth and breadth, see how others solved similar problems, and learn ideas that later shape tools, languages, and systems. It also recommends a three-pass method for reading papers, starting with the title and abstract and ending with a closer comparison to related work.

Reading notes

#

• The article frames research papers as a source of the foundations of modern software engineering and says they can change how developers approach software development.
• It groups recommended papers into system design, distributed systems, data storage and processing, metrics, modern infrastructure, computer architecture and systems performance, and search and information retrieval.
• It says reading papers builds critical thinking by showing how similar problems were handled before and by reducing the need to reinvent solutions.
• It points to foundational work on large language models as an example of research that shaped later tools like ChatGPT.
• Parnas’s paper on modularization is presented as a basis for flexibility, comprehensibility, shorter development time, and later ideas in software architecture, microservices, and API development.
• Hoare’s axiomatic basis for programming is described as a foundation for reasoning about program state and output, and for modern verification tools and type systems.
• The article says complexity in software systems needs careful study and uses the example of Out of the Tar Pit as a paper about the causes and effects of complexity.
• Hughes’s paper on functional programming is presented as an argument for modularization in functional programming.
• Lamport’s work on time, clocks, and event ordering is described as fundamental to distributed databases, blockchain, and cloud computing.
• The paper on distributed computing by Waldo and others is cited for challenging the idea that distribution disappears once a system is built.
• The Google File System is described as a scalable, fault-tolerant, high-performance distributed file system for large data-intensive applications.
• Amazon Dynamo is presented as a fast NoSQL key-value database designed for write-intensive storage, with limits and scaling considerations.
• Bigtable is described as a distributed storage system for massive structured data used across Google services.
• Codd’s relational model is presented as the theoretical foundation for SQL databases.
• MapReduce is described as a programming model for processing and generating large data sets and as a basis for modern big data frameworks.
• The metrics paper by Chidamber and Kemerer is presented as a new suite of software metrics for object-oriented design and software quality.
• Kafka is described as a distributed messaging system for high-volume log data with low latency, important for event-driven architectures.
• The Facebook memcache paper is presented as a way to scale a distributed key-value store for a very large social network.
• Bitcoin is described as a peer-to-peer electronic cash system that removes intermediaries and helps explain blockchain and decentralized systems.
• Drepper’s memory paper is presented as a bridge between hardware architecture and software development, focused on memory hierarchy, caching, and program performance.
• The Google search paper is described as introducing PageRank and the original search architecture for crawling and indexing billions of web pages.
• The article recommends a three-pass reading method: first read the title, abstract, headings, conclusions, and references; second read carefully and take notes; third try to reimplement ideas and challenge assumptions.