gantt
title Literate Programming Historical Events
dateFormat YYYY
axisFormat %Y
section Conceptual
Knuth :a1, 1984, 7y
Rossini :2001, 7y
section Typesetting
TeX78 :1978, 7y
METAFONT :1979, 7y
TeX :1982, 7y
LaTeX :1984, 7y
section Document Type
html :1991, 7y
pdf (Adobe) :1993, 15y
pdf (open) :2008, 17y
section Literate Programming
WEB :1981, 7y
Noweb :1994, 7y
section Markdown
markdown :2004, 7y
rMarkdown :2014, 7y
Quarto :2022, 3y
Appendix A — History
The interference needed to change an individual’s habits comes at a high cost, and good tools are critical for encouraging people to switch to a more literate style (Rossini 2001)
Knuth (1984a) developed \(\TeX\) in 1978 along with METAFONT with a primary purpose of improving the formatting of mathematical expressions in his books. \(\TeX\) defined spacing and layout whereas METAFONT provided fonts to use with \(\TeX\). Lamport (1985) developed \(\LaTeX\), with a user’s manual appearing in 1986. Both of these are considered typesetting systems – \(\LaTeX\) provides macros that make \(\TeX\) a bit more user-friendly for full document preparation. Both Knuth and Lamport were associated with the Stanford Research Institute. WEB was introduced in 1981 and considered a literate programming system (facilitating the writing of both computer code and documentation). WEB consists of two tools: TANGLE and WEAVE. NoWeb is also an early literate programming tool for statisticians (Ramsey 1994), and is perhaps the origin of the term chunk (as in documentation chunk or code chunk). Noweb was inspired by WEB, but had no language specificity (WEB was intended to produce either Pascal [WEB] or C code [CWEB]). NoWeb had similar tools: notangle and noweave. Rossini (2001) coined the term “Literate Statistical Practice” – his background being statistics, whereas Knuth is a computer scientist.
Stodden, Leisch, and Peng (2014) cite Sweave as an early implementation of Knuth (1984b)’s initial principles. Yihui Xie built upon the Sweave framework for knitr.
J.J. Allaire and Hadley Wickham noted their corporate charter at the Posit rebrand, with an forward-focused emphasis placed on technical communication (Allaire and Wickham 2022, see Video A.1)
A.1 Definitions
Literate Programming is a methodology for writing computer programs where the focus is on explaining to human beings what the program does rather than just instructing the computer.
Literate Programming Principles are substantive elements of computer science-derived philosophy that can be applied to domains beyond computer science. For data science, these principles include… (TBD)
\(\TeX\) a “language” (Knuth 1984a) for document formatting
Control Sequence – \(\TeX\) command consisting of escape character (\(\backslash\)) and command (the text after the escape character) that permit the printing of something that is not otherwise represented on your keyboard
Markdown is a lightweight markup language with plain text formatting syntax. It’s designed to be easily readable in its raw form and convertible to HTML, PDF, and other formats.
RMarkdown extends Markdown by allowing the integration of code (via chunks) into Markdown documents.
rMarkdown is an R package that provides tools for navigating RMarkdown documents.
Platform
1984: Knuth introduces the concept of Literate Programming in his article published in The Computer Journal (Knuth 1984b). He describes it as a paradigm where programs are written with the intent of being literature, where the explanation is as important as the code itself.
1986: Knuth’s approach gains recognition when Jon Bentley writes about literate programming in his “Programming Pearls” column in Communications of the ACM, further popularizing the concept among programmers. Bentley’s columns included examples where Knuth applied literate programming to solve real-world problems, emphasizing its benefits in clarity and maintainability.
1992: Knuth publishes “Literate Programming” as part of the CSLI Lecture Notes series, consolidating his earlier works and ideas on the subject. This book includes an anthology of his essays and serves as a comprehensive guide to the methodology, including practical examples from his work on TeX.
1995-1996: Two notable books demonstrate the application of literate programming: + 1995: Retargetable C Compiler, A: Design and Implementation by David R. Hanson and Christopher Fraser uses a literate programming approach to explain the design and implementation of a compiler, although it was more of a post hoc documentation. + 1996: C Interfaces and Implementations: Techniques for Creating Reusable Software by David R. Hanson is often cited as an example of literate programming, despite criticisms about the simplicity of its code chunk names.
2010s: There’s a resurgence of the literate programming idea with the advent of computational notebooks like Jupyter Notebooks. These tools allow for the integration of code, narrative text, and multimedia in a single document, embodying many of Knuth’s original principles but adapted for modern scientific computing and data science. This has brought literate programming closer to mainstream use, particularly in education and research settings.
2020s: The concept continues to evolve, with tools like RMarkdown and Org-mode in Emacs providing environments where literate programming is not just possible but practical for everyday use. These tools focus on reproducibility and documentation, aligning with Knuth’s vision but tailored for contemporary needs in data analysis and software development.