Writing "Learning LaTeX"
The following is a short reflection from the authors of Learning LaTeX, which was first published by SIAM in 1997. The second edition appeared in 2016.
This piece is the second installment of a new SIAM News feature called “From the SIAM Bookshelf,” which will periodically spotlight SIAM texts in areas of wide appeal to the greater applied mathematics and computational science communities.
In 1997, we published the first edition of Learning LaTeX to provide a useful resource for graduate students, early-career researchers, and other LaTeX beginners. We felt that there was a niche within the prevailing literature for a short and punchy introductory text that covered essential material while avoiding unnecessary detail. The internet was not all-pervading at the time, and the most readily available LaTeX information was present only in comprehensive, expert-oriented textbooks. We aimed for conciseness, accessibility, and concreteness in our book, which displays raw LaTeX code side-by-side with the output.
Producing Learning LaTeX was a welcome change from the more restrictive format of mathematical writing. Since we were not competing with the excellent (but somewhat dry) existing LaTeX reference books, we could afford to be highly selective, occasionally incomplete, and lighthearted — all while attempting to avoid misleading our readers and encouraging good style (\ldots instead of ...., \cos instead of cos, and \langle ..\rangle instead of <..>).
At merely 90 pages long, our inexpensive little text proved to be a popular member of SIAM’s portfolio and continued to attract interest well into the internet era, even when multiple online tutorials, beginner guides, symbol lists, and template documents became easily accessible. Many commentators have noted that humans have a penchant for the analogue, as markets still exist for paperbacks, board games, vinyl records, notebooks, and instant cameras. So perhaps a brief hardcopy LaTeX guide retains some value, even when the system generates documents that will only ever live online; for many users, flipping through a short manual in search of an example is likely more satisfying than opening yet another window in their browser. Moreover, because we focused on the absolute basics, the material was less likely to go out of date. For example, Learning LaTeX highlighted the difference between “old” LaTeX2.09 and “new, but intermediate” LaTeX\(2_\varepsilon\), which was to be superseded by LaTeX3. But so far, LaTeX3 has yet to emerge. The book’s contents thus aged less rapidly than we envisaged.
Over the years, we resisted several attempts by SIAM’s book acquisitions team to persuade us to write a new edition. They finally softened us up by highlighting a review in the esteemed TUGboat journal (the Communications of the TeX Users Group), which was written 17 years after Learning LaTeX’s initial publication. “I was surprised to find that the book is still in print,” Boris Veytsman wrote in 2013. “Thus I ordered a copy for myself and read it.” He later added that “it is a testament to (La)TeX and the authors’ efforts that this book can still be used as a first LaTeX book or as a basis for a short practical course. It would be also interesting to see a book using the same approach but updated with the new material. Making it as concise and easy to read as this one will be a challenge.”
On this basis, we signed the contract for a new edition and began preparing an update. Keeping a keen eye on the page count, we added significant new material on packages made available by the American Mathematical Society, including support for typesetting mathematical symbols and multi-line displays. We also included information on the BibTeX program for bibliography generation, the beamer class for presentation creation, and the a0poster class for poster formation. Figure 1 offers an extract from the material in section 3.5 on equation environments, which provides a sense of the text’s content and style.
The second edition of Learning LaTeX published in 2016. It clocks in at 103 pages, including a generous index (with the entries \index{recursion|see{recursion}} and \index{Welshman!bearded}). To maintain the humor-to-content ratio of the first edition, we added an imaginary list of highlights of “LaTeX Through the Years.” In the same spirit, we finish here with our predictions for future LaTeX breakthroughs, leading up to the release of LaTeX3 in 2051.
2021: New government legislation on workplace behavior bans \prod, \squeeze, and \sin.
2022: Under current social distancing rules, the second component of every susceptible-infectious-recovered (SIR) model must be typeset as dI/dt = \beta I \qquad S – \gamma I.
2023: Ph.D. student assumes that \boldmath was designed for theorems with questionable proofs.
2024: An infinite number of monkeys composing LaTeX documents fail to produce a figure in the desired location.
2025: Latest world record for “Smallest Change Necessary to Meet a Journal Page Count Requirement” stands at \vspace{–0.00001in}.
2026: Experiments show that the frequency of LaTeX errors is inversely proportional to the time remaining before a grant proposal deadline.
2027: Article disappears in a puff of smoke after mathematician types \not \exists.
2028: LaTeX’s table-positioning algorithm passes George Marsaglia’s diehard battery of statistical tests for measuring the quality of a random number generator.
2029: With a single click, Overleaf allows users to submit to a journal, complain about slow refereeing after two weeks, complain about bad refereeing after three months, and automatically resubmit to a lower-ranked journal.
2030: “Least Parsable LaTeX Command” medal awarded to \righthyphenmin.
2031: Undergraduates think that \mho will convert a statement into what they actually mean.
2032: Campus barber offers a LaTeX special: \lmoustache, \rmoustache, \curlywedge, \sqcap, and \flat\top.
2033: \newcommand{BP}{\HUGE +} turns out to be a big plus.
2034: \antihistamine introduced to combat LaTeX allergies.
2035: \search hunts down runaway arguments.
2036: A 20-page LaTeX document violates several laws of physics when it shrinks to three pages in twocolumn format.
2037: 10-year anniversary of the last recorded instance of a journal-ready article that fails to produce an “Overfull \hbox” warning.
2038: Lonely \item rescued and hand-reared by \gather, \protect, and \heartsuit.
2039: Many LaTeX users agree that \argh should be used after making an error.
2040: New LaTeX warning implemented: “Paragraph ended before you wrote anything of any interest.”
2041: Plans to reintroduce \smallskips back into the wild are scuppered by the predatory instincts of the non-indigenous \vfill population.
2042: Tiebreaker question in final of “International Copyeditor of the Year” asks contestants to differentiate between the results of f: \mathbb{R} \to \mathbb{R} and the correct, but almost indistinguishable, f \colon \mathbb{R} \to \mathbb{R}.
2043: TeXpert demoted to rank of TeXnician after using Microsoft Word in public.
2044: \fussy and \sloppy file for divorce, citing irreconcilable differences.
2045: \supset and \frown become increasingly irritated by \smile.
2046: In an effort to curb record levels of font inflation in the U.K., Bank of England declares that \footnotesize is the new \normalsize.
2047: A rogue \partial /\partial d x escapes from a computer laboratory, wreaking havoc within the local x^k community.
2048: Survey reveals that 87 percent of LaTeX users import lengthy preambles, donated by well-meaning friends, that nobody has completely understood since 1993.
2049: Worldwide mathematical productivity increases dramatically after discovery of a Newton method alternative to the default troubleshooting technique of bisection with \end{document}.
2050: BibTeXing of arXiv articles remains a mystery to all.
Enjoy this article? Visit the SIAM bookstore to learn more about Learning LaTeX and browse other SIAM titles.
About the Authors
David F. Griffiths
Retired Reader, University of Dundee
David F. Griffiths is a retired Reader in Applied Mathematics at the University of Dundee.
Desmond J. Higham
Professor, University of Edinburgh
Desmond J. Higham is a professor of numerical analysis at the University of Edinburgh. He is a SIAM Fellow.