(in progress of restructuring the guide)
The recommended way is to add this project as a submodule of your git project. In shell, enter the directory of your git respository, and type
git submodule add git@github.com:uw-pluverse/pluverse-latex-style-guide.git
Then in your main.tex
, type the following text to use the pluverse preamble files,
where constants.tex
is a file in your project that stores all your project-specific
constants.
\input{pluverse-latex-style-guide/macro/pluverse-preamble.tex}
\input{constants}
\input{pluverse-latex-style-guide/macro/pluverse-preamble-end.tex}
To clone a Git repository along with all its submodules,
you can use the --recurse-submodules
flag with the git clone
command.
This ensures that all submodules are initialized
and updated to match the versions specified in the superproject.
Here's the command:
git clone --recurse-submodules <repository_url>
Replace <repository_url>
with the URL of the repository you want to clone.
If you have already cloned a repository without its submodules, you can initialize and update the submodules separately using the following commands:
git submodule update --init --recursive
This will initialize, fetch, and checkout the submodules.
Update the submodules to the latest commit from the upstream repository (if the submodule repository has new commits):
git submodule update --remote --merge
-
Git is much better at tracking changes than overleaf.
-
You can use TexStudio, vim or other editors locally and push you changes to the remote git repository.
If you are using Texstudio, please follow this: https://tex.stackexchange.com/a/282571.
Keep a line short. I would recommend to break a line at 60. The benefits of doing this are
- Friendly to version control system. Easy to diff.
- Better bi-directional mapping between the output pdf file and the latex source.
- Friendly to the editor. No need to use the editor's wrapping functionality.
- Friendly to version control system. Easy to diff. Minimize merge conflicts.
- Better organization of the paper, for example, each section is in its own file.
- Easy to identify the entry of the tex project.
- Enter your latex directory, and run the following command
$ wget https://raw.githubusercontent.com/uw-pluverse/pluverse-latex-style-guide/master/macro/update-pluverse-macros.sh
$ chmod +x update-pluverse-macros.sh
$ ./update-pluverse-macros.sh
Use this website to check whether your title conforms. Note that this rule applies to paper titles, subtitles, and all (sub)section titles.
- Purpose: This command is used to emphasize text. It is semantically meaningful and is intended to emphasize text in a context-sensitive way.
- Usage: The text inside the braces {} will be emphasized.
- Behavior: The \emph command is context-sensitive. If used inside already emphasized text (e.g., inside another \emph), it will typically de-emphasize the text (e.g., change it back to the normal style).
This allows the citation or the figure label to stay together with the previous word with a whitespace in between.
A and B developed a new approach~\cite{paper-id}
Figure~\ref{figure-label} shows the overall framework.
Correct: In Figure 1, we know ...
Wrong: In figure 1, we know ...
- Table: \label{tbl:...}
- Subtable: \label{subtbl:...}
- Figure: \label{fig:...}
- Subfigure: \label{subfig:...}
- Section: \label{sec:...}
- Subsection: \label{subsec:...}
- Subsubsection: \label{subsubsec:...}
- Algorithm: \label{alg:...}
- Line: \label{line:...}
This applies to most of the proceeding formats.
The first word in caption of Table/Figure should be capitalized, but not for preposition and conjunction words.
For example, it should be The Effectiveness of ADF in Three Datasets
.
Here is a detailed guideline from the feedback from a proceeding publisher:
Headline-style capitalization should be used.
Capitalize:
- first and last word, first word after a colon
- all major words (nouns, pronouns, verbs, adjectives, adverbs)
Lowercase:
- articles (the, a, an)
- prepositions (regardless of length)
- conjunctions (and, but, for, or, nor)
- to, as
Note:
- The word immediately preceding the hyphen should be capitalized. For example, it should be
Learning-Based Approach
. An exception isX-ray
, instead ofX-Ray
.
-
Try to avoid using
.png
,.jpg
and.bmp
, because they look blury when zoomed in. -
Always use
.eps
or.pdf
. -
Make sure the font in the figure is the same font.
The color and markers of figures, especially for the line chart and bar chart, should be able to distinguished in black-white printing.
These three websites can help you pick colors:
Remember to pick the color that is colorblindness-friendly and print-friendly.
Usually it should be Times New Roman.
By default, the font of the figures generated by Python matplotlib is Type-3 font, which is not accepted by the checking system of camera ready. Use the following commands in Python to avoid this.
import matplotlib
matplotlib.rcParams['pdf.fonttype'] = 42
matplotlib.rcParams['ps.fonttype'] = 42
Be carefully when you use resize.
For table, you can use this website https://www.tablesgenerator.com/. Now it is more common to use the "booktab table style".
Here is a guide on how to design good-looking tables in LaTex by Markus Püschel.
Use \setlength
and \arraystretch
to change the horizontal spacing (column separation) and the vertical spacing (row separation), respectively.
\setlength{\tabcolsep}{20pt}
\renewcommand{\arraystretch}{1.5}
If you use numbers such as 1
, 2
, 100
, and 1000
, you should write one
, two
, one hundred
, and one thousand
.
If you use number such as 2021
, it should write it as 2,021
More information can be found here
Do not use English word abbreviations, such as 2nd
, no.
and approx.
. Use the full words instead, such as second
, number
and approximately
. You should also avoid using they're
, don't
, can't
, etc; instead you should use they are
, do not
, cannot
.
Note that you are encouraged to use Latin abbreviations, such as i.e.
, e.g.
.
A benchmark is a test that measures the performance of hardware or software. For example, in the evaluation of Perses, each subject (e.g., clang-22382) can be considered as a benchmark. A set that includes multiple benchmarks is called a benchmark suite.
Reference: SPEC CPU2006 and SPEC CPU2007
- Ratio
- A ratio says how much of one thing there is compared to another thing. It shows that the number of times one value contains or is contained within the other.
- Formula:
$A/B$ - Part-to-Part vs Part-to-Whole
- Part-to-Part: comparing one part to another part
- e.g., ratio of bug discovered to bug missed = 2 : 3 or 2/3
- Part-to-Whole: comparing one part to the whole lot
- e.g., ratio bug discovered to all bugs = 2 : 5 or 2/5
- Part-to-Part: comparing one part to another part
- E.g., Perses is 2.07x faster, produces 1.13x smaller results, and reduces 3.99x fewer queries than HDD.
- link
- Percentage change / Relative change / Relative difference
- This shows that change as a percent of the old value. When the new value is greater then the old value, it is a percentage increase, otherwise it is a decrease.
- Watchout:
- A 10% increase from 100 is an increase of 10. BUT! A 10% reduction from 110 is a reduction of 11.
- Formula:
$(New Value − Old Value) / Old Value \times100%$ - E.g., Perses’s results are respectively 2% and 45% in size of those from DD and HDD.
- Acronym: %Change
- link
- Percentage error
- The difference between Approximate and Exact Values, as a percentage of the Exact Value. Percentage Error is all about comparing a guess or estimate to an exact value.
- Formula:
$(Approximate Value − Exact Value) / Exact Value \times100%$ - E.g., A bug detecting tool only found 95 bugs out of a 100. The tool has a 0.05% error.
- Acronym: %Error
- link
- Percentage difference
- The difference between two values divided by the average of the two values. Percentage Difference is used when both values mean the same kind of thing (for example the heights of two people).
- Formula:
$|First Value − Second Value| / |(First Value + Second Value)/2| * 100%$ - Acronym: %Diff or %Difference
- link
- Ratio: how much of one value is compared to another value
- Percentage change: compare an Old Value to a New Value
- Percentage error: compare an Approximate Value to an Exact Value
- Percentage difference: both values mean the Same Kind of thing (rarely used)
- Commonly, we prefer using ratio and %change in our evaluations. %error and %diff is used less frequently as they only apply to specific scenarios.
- Also see this link
Please try to download the paper and bib from ACM Digitial Library and IEEE Xplore.
DBLP is a good source for tracking a person's publication records and download bib files. Just be careful: sometimes a paper can appear twice if it is published both in conference/journal and arxiv. Please DO NOT select the one from arxiv.
Avoid the citation from arxiv, if possible.
You should not cite any papers in the Abstract
section.
If you have some numbers, or string literals (e.g., functions) used throughout your paper, create a macro, and use that macro instead.
For example, if you repeatedly mention Our approach is 35% faster than the state of the art.
, then it is better to refactor your LaTeX into the following
usepackage{xspace}
\newcommand{\PerformanceGain}{35\%\xspace}
......
Our approach is \PerformanceGain faster than the state of the art.
You can also use Latex to compute simple arithmetics.
\newcommand{\cSizeDecRateProjVsPerses}{%
\pgfmathparse{(1 - \cSizeAvgProjValue/\cSizeAvgPersesValue)*100}
\pgfmathprintnumber{\pgfmathresult}\%\xspace%
}
Single quotation marks are produced in LaTeX using ` and '.
Double quotation marks are `` and ''.
For example
`Something'
``The title of this paper is hope''
Trailing whitespaces are not friendly to git, because it introduce unnecessary diff.
If you use TeXStudio, you can enable removing trailing whitespaces on save in the setting
.
- On MacOs, From the top menu, click
TeXStudio->Preference->Editor
, and tickRemove Trailing Whitespace on Save
- Not sure how to set this on other OS.
Use \S\ref{section-label}
. Note that there is no space between \S
and \ref{section-label}
If you want to use words in a math formula, remember to use \textit to wrap the word.
For example, $time = ComptueTime(t + 1)$
should be written as $\textit{time} = \textit{ComputeTime}(t + 1)$
ACKNOWLEDGMENTS section shall not have a section number. It should be added using the environment \begin{ack} \end{ack}
instead of \section{ack}
Please note it is ACKNOWLEDGMENTS, not ACKNOWLEDGEMENTS.
\section{Approach}
ADD SOMETHING HERE!
\subsection{Overview}
TBD: need to have an example here.
Cleveref automatically determine the type of cross-reference and fill in the appropate cross-reference (Figure, Table, Section, etc.).
To use cleveref, load the cleveref package last and use the following configurations.
\usepackage{cleveref} % This package must be loaded in the end.
\Crefname{algocf}{Algorithm}{Algorithms}
\crefname{algocf}{Algorithm}{Algorithms}
\Crefname{algorithm}{Algorithm}{Algorithms}
\crefname{algorithm}{Algorithm}{Algorithms}
\Crefname{figure}{Figure}{Figures}
\crefname{figure}{Figure}{Figures}
\crefname{listing}{Listing}{Listings}
\Crefname{listing}{Listing}{Listings}
\Crefname{table}{Table}{Tables}
\crefname{table}{Table}{Tables}
\crefname{thm}{Theorem}{Theorems}
\Crefname{thm}{Theorem}{Theorems}
% https://tex.stackexchange.com/questions/81634/cleveref-configure-symbol-for-all-sectioning-types-once-time
\crefformat{chapter}{\S#2#1#3}
\crefmultiformat{chapter}{\S\S#2#1#3}{ and~#2#1#3}{, #2#1#3}{, and~#2#1#3}
\crefformat{section}{\S#2#1#3}
\crefmultiformat{section}{\S\S#2#1#3}{ and~#2#1#3}{, #2#1#3}{, and~#2#1#3}
When referencing labels, instead of using \ref, use \cref.
- In a defining clause, use that.
- In non-defining clauses, use which.
Less
is sometimes used with a plural countable noun (e.g., less cars), particularly in conversation. However, this is grammatically incorrect.
Note that "cf." is not the appropriate abbreviation for referring to details in a specific section. "cf." stands for "confer" in Latin, meaning "compare" or "consult," and is used to direct the reader to other material for comparison, not for additional details.
For example,
ME:
Which version is grammatically correct?
- Company A has a larger market share, compared to Company B.
- Company A has a large market share, compared to Company B.
ChatGPT:
The first sentence is grammatically correct:
Company A has a larger market share, compared to Company B.
This sentence correctly uses the comparative form "larger" to compare the market shares of Company A and Company B. The second sentence, "Company A has a large market share, compared to Company B," is not incorrect, but it is less precise because it does not explicitly indicate the comparative nature of the market shares.
Try
is an informal word.
For example, very effective
can be replaced with highly effective
.
You might also consider remarkably effective
or amazingly effective
, though Chengnian does not like them much.
From the grammarly post
Although lots and plenty are acceptable in academic writing, their usage is considered to be informal. In formal academic writing, it is more appropriate to use many, much, and more.
Correct: X......, etc.
Incorrect: X......., etc..
When a noun is countable and singular, you need to use a/an/the
to quantify it.
These two phrases cannot be used together. For example, such as apple, organge, etc.
.
If you use such as
, then you should strictly follow the structure such as A, B, C and D
.
For example
ddmin can handle ...;
In stead, you should write
The algorithm ddmin can handle ...;
Similarly, instead of writing 1 apple a day is good for your body
, you should write One apple a day ...
.
"have no clue", reason: informal idiom, see
comment_macror.tex
, to leave commentsauto_counter_macro.tex
, If you have a lot of findings/reponses, you can use the following commands to automatic generate counters.abbr_macros.tex
, macro for common macros.response_letter.tex
, a rough template for journal response letter.
Use \protect
before the function macro, e.g., \caption{\protect\Fn my function}
See here
The pluverse-preamble.tex
file provides commenting macros, such as \cn{}
. When using these macros, ensure that each macro begins on a new line and that the closing parenthesis is followed by another new line. If your comment spans multiple lines, make sure to use indentation within the macro for clarity.
For example, the following adheres to the recommended style:
\cn{This is a one-line comment.}
\cn{This is a multi-line comment.
line 2 % Note the indentation here to highlight the current line is a comment.
line 3 % Note the indentation here to highlight the current line is a comment.
}
We recommend this style because it makes it easier to comment out or remove these macros. For instance, in the following LaTeX code, it is difficult to remove the comment macros cleanly:
Main text here. \cn{My comment is embedded here, and it is not easy to delete.} Main text here.
This example demonstrates a similar issue:
Main text here. \cn{My comment spans two lines
} Main text here.
If you receive a comment from a reviewer and you address it, then do not delete the comment but leave another comment \YourCommentMacro{done}. Doing this can notify your reviewer of your response. If the reviewer feels that their question is well addressed, the reviewer will comment out the macro.
For example, assume I am a reviewer and leave the following comment
\cn{My comment here.}
Then you address my comment and leave a done
comment
\cn{My comment here.}
\victor{done}
If I feel my question is addressed, I will simply comment out all comment macros like the following
%\cn{My comment here.}
%\victor{done}