Where the ideas were first written down
Founding Papers
Before a technology becomes a product, someone argues for it on paper. These are the short documents behind the essays on this site, read the way a strategist would read them: what problem the author faced, what they bet on, and how the bet looks now.
- 1945 As We May Think The record of human knowledge has outgrown the ways we have to search it, and a machine that follows the mind's own associations could fix that.
- 1960 Man-Computer Symbiosis The most productive near-term future is a close partnership in which people set goals and judge, and computers do the routine work, which requires interactive, shared computers rather than batch processing.
- 1962 Augmenting Human Intellect: A Conceptual Framework Human intellectual capability comes from a system of the person, their language, their tools and their methods, together with their training, and improving any part changes the whole, so computers should be designed to co-evolve with how people think and work.
- 1965 Cramming More Components onto Integrated Circuits The number of components on the cheapest-to-make chip has been doubling roughly every year, and that rate should continue for at least a decade, which will make electronics dramatically cheaper and more pervasive.
- 1968 Go To Statement Considered Harmful Because people are much better at understanding static structures than processes unfolding in time, programs should be written so that their text mirrors their execution, which means abandoning the unrestricted go to statement.
- 1970 A Relational Model of Data for Large Shared Data Banks Data should be presented to users as simple mathematical relations, tables, with no reference to how it is physically stored, and queried by a high-level language, which makes applications independent of storage and gives a rigorous basis for consistency.
- 1974 The UNIX Time-Sharing System A simple, coherent system, organised around a hierarchical file system and small composable programs, can match far larger systems in usefulness at a fraction of the cost, and constraints of size encourage elegance.
- 1984 End-to-End Arguments in System Design A function that can only be completed correctly by the applications at the ends of a communication should be implemented there; putting it in the lower layers of the network is at best a performance optimisation and often a waste.
- 1986 No Silver Bullet: Essence and Accidents of Software Engineering No single development will give an order-of-magnitude improvement within a decade, because most of the remaining difficulty of software lies in its essence, the conceptual work of specifying and designing complex systems, rather than in accidental difficulties of tools and representation.
- 1989 Information Management: A Proposal Information in an organisation is a web, not a tree, so it should be managed as a distributed system of linked notes that anyone can add to, with no central control.