Technical Hints for Writing Papers

This chapter is based on the notes 2022-07-19_Evaluation_PaperWriting.pdf

Paper Structure

• Abstract

• Introduction + Related Work

• Concepts

• Evaluation

• Conclusion

• Statements (Data Availability, Funding, Acknowledgements)

• References

Paper Content

• See review criteria

• Tell the whole story in the introduction, rest is detail

Methodology for Paper Writing

P.R. Halmos:

There is no recipe and what it is.

The basic problem in writing mathematics is the same as in writing biology, writing a novel, or writing directions for assembling a harpsichord: the problem is to communicate an idea.

To do so, and to do it clearly, you must…

• have something to say,

• have someone to say it to,

• organize what you want to say,

• arrange it in the order you it said in,

• write it, rewrite it, and re-rewrite it several times,

• be willing to think hard about and

• work hard on mechanical details such as diction, notation, and punctuation.

That’s all there is to it.

Psychological tricks:

• empty-page effect

• divide and conquer

• small steps, incremental

• start, even if you think the outcome will be marginal

Approaches to writing:

• Treat writing a paper like programming/software development

• Source code is precious, should be of high quality and maintenable

• Consider teamwork / shared code

Consequence: Source code of paper should be:

• well readable: structured, organized

• minimal: refactored, avoid repetitions with macros, remove unnecessary comments

Examples:

• Do not write ‘CPAchecker’, but use \newcommand{\cpachecker}{CPAchecker}. If you change typesetting to \textsc{CPAchecker}, only single location must be changed

• Do not print numbers of experimental data in table and repeat them in text, but use \newcommand{\correctAlarmsCpachecker}{238}. Avoids typos and inconsistencies (also makes automation easier).

Source: Our wiki

More sources:

• Armando Fox: Armando’s Paper Writing and Presentations Page

• Donald E. Knuth, Tracy Larrabee and Paul M. Roberts: Mathematical Writing

• Joel E. Cohen: To A Young Scientist

• Mary-Claire van Leunen and Richard Lipton: How to Have Your Abstract Rejected

• Mark Wegman: What it’s like to be a POPL referee; or how to write an extended abstract so that it is more likely to be accepted

• P. R. Halmos: How to Write Mathematics

• Simon Peyton Jones: How to write a good research paper, give a good research talk, and write a good grant proposal

• William Pugh: Advice to Authors of Extended Abstracts

Text editing

• (at least) every sentence on new line

• avoid changing line breaks (do not auto-wrap lines)

• use short lines

• avoid tabs, use spaces

• commit everything to repository that is required to build the project

• also commit the bibliography file (main.bbl)!

• always check diff before push

Example diff to show how short lines are helpful: What is the actual change?

- To increase the reliability of these results, validators like \textsc{CPAchecker} \cite{CPACHECKER}, \textsc{UAutomizer} \cite{UAUTOMIZER2013}, and \textsc{MetaVal} \cite{MetaVal} use witnesses to reverify programs wtih desirably less effort by trying to reduce the state space by strengthening the used analysis with
+ To increase the reliability of these results, validators like \textsc{CPAchecker} \cite{CPACHECKER}, \textsc{UAutomizer} \cite{UAUTOMIZER2013}, and \textsc{MetaVal} \cite{MetaVal} use witnesses to reverify programs with desirably less effort by trying to reduce the state space by strengthening the used analysis with

vs.

To increase the reliability of these results,
validators like \cpachecker~\cite{CPACHECKER},
\uautomizer~\cite{UAUTOMIZER2013}, and \metaval~\cite{MetaVal} 
use witnesses to reverify programs
- wtih desirably less effort
+ with desirably less effort
by trying
to reduce the state space by strengthening the used analysis
with

Comments

• % Text that may be used in the future

• %%% db: Explanation by Dirk Beyer for co-authors

• Each comment starts a new line

Macros

• \newcommand{..}[#args]{..}

• Never use \def! Silently overwrites existing macros

Writing Style

• avoid ambiguity with symbols: ‘a set S of states’, not ‘a set of states S’. With the latter, S could also be a state.

• ‘iff’ in conditions (theorems), ‘if’ in definitions

Spaces and Periods

• french spacing: ‘D. Beyer’ -> D. Beyer (single space, not larger)

• paragraphs: \noindent, \smallskip, \medskip, \bigskip`

• hair space between quantity and unit: 33\,GB (best solved with siunitx)

• use short symbols: ‘the state~s’, ‘for all~i’

• citations:

  • Reference is not part of sentence, never used as object in sentence. Removing the reference may not destroy the sentence. Bad: “[3] says …”, “as described in [4]” Good: “as described by Smith [4].”, “as done in HCI [5]”

  • Always use non-breaking space ~\cite{...}

  • Avoid “et al.”

Types of Hyphens and Dashes

• Hyphen: Conjuncts words

• En-dash: Minus and range

• Em-dash: Separation of thoughts, insertion

BibTeX

• Use proper names for entry keys

  • R-textbook

  • Fool-Name, Author-Name + Year

  • Conference

  • Best: Memorizable abbreviation

• Entry data should be in consistent style and formatting

• Use ‘annot’ field for short explanation why paper is relevant

Repository Organization

• Start repository as soon as serious work on topic exists (e.g., whiteboard screenshot)

• Commit each document

• Main source file: main.tex

• Split large documents into sections. For example: main.tex, abstract.tex, intro.tex, …

• Sub directories:

  • archive/

    • 2021-03-02_CAV07-submission.pdf

    • 2021-07-05_CAV07-published.pdf

    • 2021-07-05_CAV07-Copyright-Form.pdf

    • 2021-05-05_CAV07-rebuttal-answer.txt

  • paper/

    • main.tex

    • …

  • talk/

    • main.tex

    • …