DEV Community: Saurabh Shah

LaTeX for Academic Publishing: What Journals Actually Require (And What's Optional)

Saurabh Shah — Tue, 24 Mar 2026 02:37:00 +0000

A practical guide based on real journal submission experience - not the textbook version.

"Submit in LaTeX" covers a lot of ground. Some journals require it strictly. Others accept Word but strongly prefer LaTeX. A few say "LaTeX preferred" and mean "we'll accept either but we'll reformat your Word document and it will look different from what you submitted."

After handling LaTeX formatting for academic papers across IEEE, ACM, Springer, Elsevier, MDPI, and arXiv at The LaTeX Lab, here's what we actually know about what different publishers require, where submissions break, and what matters most.

Which Publishers Actually Require LaTeX

Effectively mandatory:

IEEE - all Transactions, Letters, and conference proceedings. IEEEtran is the required class. Submitting Word gets your paper reformatted by the editorial office, often with formatting that doesn't match what you intended.
ACM - all venues using the acmart class (SIGCHI, SIGCOMM, PLDI, etc.). Their submission system accepts PDFs but the camera-ready requires the LaTeX source.
arXiv - requires LaTeX source for most submissions. PDFs without source are accepted but flagged as lower confidence for reproduction.
Most mathematics journals (AMS, SIAM) - LaTeX is the de facto standard; Word submission is unusual enough that editorial staff often struggle with it.

Strong preference, not technically required:

Springer - LNCS conference proceedings require LaTeX. Journal articles accept Word but the typesetting quality is noticeably different.
Elsevier - accepts Word for most journals but the elsarticle LaTeX class is the production format. Authors who submit LaTeX go through less reformatting in production.

Genuinely either/or:

MDPI - accepts both. Their mdpi.cls template is well-maintained and produces clean output, but Word submissions are handled with similar quality.
PLOS ONE - accepts Word or LaTeX. No strong preference either way.

The Three Most Common Journal Template Mistakes

1. Loading packages the class file already loads

Every journal class file loads a set of packages internally. When you add them again in your preamble, you get option clashes or redefinition warnings.

The common offenders with IEEEtran:

% Don't load these separately with IEEEtran - it handles them:
\usepackage{cite}        % IEEEtran has its own citation handling
\usepackage{graphicx}    % already loaded
\usepackage{amsmath}     % load this one is fine, but check first

% Same issue with elsarticle:
\usepackage{natbib}      % elsarticle loads natbib internally

Read your class file's documentation before adding packages.

2. Wrong citation style for the journal

Every publisher has a required citation style. Getting it wrong is a desk rejection risk at some venues, and always requires a revision round.

% IEEE: numeric citations, sorted by appearance
\bibliographystyle{IEEEtran}

% ACM (acmart class): handles citation style internally via class options
\documentclass[format=acmsmall, review=false, screen=true]{acmart}

% Springer LNCS: numeric, sorted by appearance
\bibliographystyle{splncs04}

% Elsevier (author-year for most journals):
\bibliographystyle{elsarticle-harv}  % Harvard author-year
% or
\bibliographystyle{elsarticle-num}   % numeric

Check the journal's author guidelines - not the template README, the actual guidelines page. They often contradict each other.

3. Two-column float handling

IEEE and ACM use two-column layouts. Figures and tables that span both columns need figure* and table* environments. Using figure for a wide figure produces output that overflows into the second column.

% Single-column figure (fits in one column):
\begin{figure}[t]
  \includegraphics[width=\columnwidth]{fig1}
  \caption{Single-column figure.}
\end{figure}

% Full-width figure spanning both columns:
\begin{figure*}[t]
  \includegraphics[width=\textwidth]{fig2}
  \caption{Full-width figure spanning both columns.}
\end{figure*}

The * variants can only be placed at the top or bottom of a page ([t] or [b]), not inline. This causes figures to float further than expected - usually fixed by adjusting the figure's position in the source relative to where you want it to appear.

What "Template Compliance" Actually Means in Practice

Journal editorial offices run their own checks on submitted LaTeX. What they're looking for:

Margins. The class file handles this - don't override \geometry settings unless the template documentation says to. Any margin change that makes the paper shorter (easier to read for reviewers) will be caught.

Font size. Some conferences enforce 10pt strictly. \small or \footnotesize in the main body text is a flag.

Reference format. This is the most common reason for revision requests. A missing DOI field, an inconsistent journal name abbreviation, or conference papers formatted as journal articles.

Figure resolution. PDFs submitted to most venues should have figures at 300dpi minimum for raster images. Vector figures (PDF, EPS, SVG) don't have this issue - use them wherever possible.

arXiv-Specific Notes

arXiv has a few requirements that trip up first-time submitters:

All files in one directory. arXiv's compiler expects your .tex, .bib, all figure files, and your .cls/.sty files in a flat directory or a simple zip. Nested subdirectories for figures sometimes cause compilation failures.

No \pdfoutput=1. arXiv sets this itself. Including it in your preamble can cause conflicts.

BibTeX, not BibLaTeX. arXiv's compilation environment supports BibTeX with bibtex. BibLaTeX with biber is supported but less reliable. If you're targeting arXiv, test with the BibTeX backend.

The .bbl file. If arXiv can't run BibTeX on your submission, it won't compile the bibliography. Submit the pre-compiled .bbl file alongside your .tex to guarantee bibliography output.

The LaTeX Compilation Errors That Waste the Most Time (And How to Fix Them Fast)

Saurabh Shah — Mon, 23 Mar 2026 01:10:00 +0000

The errors we see most often after fixing hundreds of broken LaTeX projects - with exact fixes.

LaTeX error messages are famously unhelpful. The line number is wrong, the message describes a symptom not a cause, and the actual problem is three packages up in the preamble.

After fixing compilation errors across hundreds of academic papers and journal submissions at The LaTeX Lab, we've mapped which errors waste the most time and what actually fixes them. Here's the list.

1. `! Undefined control sequence`

The most common error in LaTeX. It means you've used a command that LaTeX doesn't recognise.

The three actual causes:

Missing package. You used \textcolor{red}{text} but forgot \usepackage{xcolor}. The fix is always in the preamble - find the package that defines the command.

% Fix: add the right package
\usepackage{xcolor}     % for \textcolor
\usepackage{amsmath}    % for \text, \align, \DeclareMathOperator
\usepackage{booktabs}   % for \toprule, \midrule, \bottomrule

Typo in command name. \beginn{itemize} instead of \begin{itemize}. Check the exact spelling.

Command defined in a class file you're not loading. Common with journal templates - you copied a snippet from a paper that uses \IEEEauthorblockN{} but you're not using IEEEtran. Either load the right class or find the equivalent command in your class.

2. `! Missing $ inserted`

LaTeX encountered a math symbol outside of math mode. Usually happens in one of three situations:

% Problem 1: Underscore in text
The variable x_1 is defined as...   % ← _ triggers math mode error

% Fix:
The variable $x_1$ is defined as...

% Problem 2: Math command in text
The value is \alpha degrees.         % ← \alpha only works in math mode

% Fix:
The value is $\alpha$ degrees.

% Problem 3: Underscore in a label or filename
\label{eq:result_final}              % ← fine in labels, but check captions
\caption{Result\_comparison}         % ← escape underscores in text contexts

The line number reported in this error is almost always wrong - LaTeX reports where it gave up, not where you made the mistake. Search backward from the reported line.

3. Package Conflicts (`Option clash for package X`)

This one is almost always caused by the same package being loaded twice with different options, either by you or by the class file.

! LaTeX Error: Option clash for package xcolor.

% The problem - you load it one way:
\usepackage[dvipsnames]{xcolor}

% The class file loads it another way internally:
\usepackage{xcolor}  % no options

% Fix - load the package BEFORE the class, or use:
\PassOptionsToPackage{dvipsnames}{xcolor}
\documentclass{IEEEtran}  % now the class inherits your options

The \PassOptionsToPackage command is the correct tool for resolving option clashes when you can't modify the class file.

4. `! File 'X.sty' not found`

A package isn't installed in your TeX distribution.

On Overleaf: this almost never happens - Overleaf has a near-complete TeX Live installation. If you see it on Overleaf, the package name is probably misspelled.

On a local install: the package needs to be installed via your TeX package manager.

# TeX Live (Linux/Mac):
tlmgr install packagename

# MiKTeX (Windows):
# Use the MiKTeX Package Manager GUI, or:
mpm --install=packagename

If tlmgr says the package doesn't exist, check CTAN - the package may have been renamed or merged into another package.

5. `Runaway argument` / `Paragraph ended before \X was complete`

This usually means an unmatched brace or a blank line inside a command that doesn't allow it.

% Common cause 1: Missing closing brace
\textbf{Some important text   % ← no closing }

% Common cause 2: Blank line inside \footnote or \caption
\caption{This is a table showing
         the results.           % ← blank line here breaks the argument

% Fix:
\caption{This is a table showing the results.}

Use your editor's brace-matching feature to find the unmatched brace. In Overleaf, the bracket highlighting will show you where the match is missing.

6. BibTeX Errors That Don't Show Up Until You've Fixed Everything Else

BibTeX errors are separate from LaTeX compilation errors and show up in the .blg log file, not the main log. Common ones:

Warning--missing journal in Jones2021
Warning--empty author in Smith2022
I was expecting a `,' or a `}'

The first two are missing required fields - BibTeX will still compile but the bibliography entry will be malformed. Fix them in your .bib file.

The third means a syntax error in your .bib file - usually a missing comma between fields or an unescaped special character:

% Problem: unescaped & in journal name
journal = {Transactions on Systems & Control}

% Fix:
journal = {Transactions on Systems \& Control}

Always check the .blg file after a full compile. LaTeX can report zero errors while your bibliography is silently malformed.

7. `Overfull \hbox` Warnings (The One You Should Actually Fix)

Overfull \hbox is a warning, not an error - LaTeX tells you text is sticking out into the margin. Most people ignore these. You shouldn't, because journals check for them.

% Common cause: a URL or long word that can't break
\url{https://verylongdomainname.com/with/a/very/long/path/to/something}

% Fix 1: allow URL line breaks
\usepackage{url}   % or hyperref, which includes url

% Fix 2: for a long word, add a discretionary hyphen:
super\-cali\-fragi\-listic  % LaTeX can break at these points

% Fix 3: use \sloppy locally for a paragraph that won't cooperate:
{\sloppy This paragraph has a very long URL that LaTeX can't break nicely. \par}

The Fastest Debug Workflow

When you have multiple errors, don't try to fix them all at once. LaTeX errors cascade - one missing brace causes 40 downstream errors. The workflow:

Fix only the first error in the log
Recompile
Repeat

Also: comment out everything except the preamble and one paragraph, confirm it compiles, then add content back in sections. Binary search is faster than reading 200 error lines.

PDF to LaTeX Conversion: Why It's Hard and What Actually Works

Saurabh Shah — Sun, 22 Mar 2026 09:30:00 +0000

Why automated PDF to LaTeX tools produce unusable output - and the right approach for academic documents.

PDF to LaTeX is a harder problem than Word to LaTeX, and it's worth understanding why before you spend time trying to automate it.

At The LaTeX Lab, PDF conversions make up a significant portion of the projects we handle - researchers who only have a final PDF of their paper, no original source file. Here's what we've learned about where the process breaks and what to do about it.

Why PDFs Are a Poor Source for LaTeX Reconstruction

A PDF is a rendering format. It stores instructions for placing glyphs on a page at precise coordinates. It does not store:

Semantic structure (what is a heading vs. body text)
Mathematical relationships (what is a fraction, what is a subscript, what is an operator)
Table structure (where rows and columns begin and end)
Bibliography metadata (author, journal, DOI - just the rendered string)

When a PDF to LaTeX converter processes a document, it's reverse-engineering rendered output back into structural markup. For plain body text, this works tolerably. For everything else, it doesn't.

What Automated PDF to LaTeX Tools Actually Produce

Take a simple display equation rendered in a PDF - say, a fraction inside a summation. In the PDF, this is a set of glyph positions: a sigma character at coordinate (x1, y1), a fraction bar at (x2, y2), numerator glyphs above it, denominator below. There's no metadata saying "this is a summation with lower limit n=0 and upper limit N."

An automated converter has to guess the mathematical structure from the spatial arrangement of glyphs. For simple equations it sometimes guesses correctly. For anything involving:

Nested fractions
Matrix environments
Aligned multi-line derivations
Custom operators or symbols
Subscripts and superscripts on top of each other

...the output ranges from wrong to completely absent. The equation either gets skipped, rendered as an image extracted from the PDF, or reconstructed incorrectly.

% What pdf2latex-style tools give you for a complex equation:
\includegraphics[width=0.8\textwidth]{eq_extracted_01.png}
% Or worse, nothing at all.

% What the equation actually needs to be:
\begin{equation}
  \hat{y} = \sum_{n=0}^{N} w_n \cdot \phi\!\left(\frac{x - \mu_n}{\sigma_n}\right)
\end{equation}

Tables: Even Harder Than Equations

Table reconstruction from PDF is genuinely unsolved as an automated problem. The PDF stores each cell's text as positioned glyphs - it has no concept of rows, columns, or cell boundaries except as inferred from whitespace.

For simple two-column tables with clear spacing, automated tools produce something usable. For tables with:

Merged cells (colspan/rowspan)
Ruled lines between specific rows
Multi-line cell content
Rotated headers

...the output is a jumbled list of strings, not a table. We've never seen an automated tool handle a longtable or a tabularx with multi-line cells correctly.

The Text Layer Problem on Scanned PDFs

If the PDF was produced by scanning a physical document, there's an additional layer: OCR. Scanned PDFs don't have a text layer at all - they're images. Any text extraction first requires OCR, which introduces its own error rate.

For academic documents with domain-specific notation, OCR error rates on equations are high. A scanned fraction is frequently OCR'd as something like H/2 when it should be \hbar / 2 - visually similar characters, semantically different.

If you're working from a scanned PDF, the realistic path is:

OCR the document with a high-accuracy engine (ABBYY FineReader or similar)
Manually verify all equations and symbols against the original
Reconstruct the LaTeX from scratch using the OCR output as a reference

There's no shortcut here that produces submission-ready output.

What the Right Approach Looks Like

For academic documents, the only reliable approach to PDF to LaTeX is semi-manual reconstruction:

Extract clean body text from the PDF (where the text layer exists)
Read every equation directly from the PDF and typeset it in LaTeX math mode from scratch
Reconstruct every table in booktabs format from scratch
Extract and rebuild the bibliography as a .bib file - either from the PDF's reference list or by looking up each reference in CrossRef/Google Scholar to get clean metadata
Apply the journal or university template and compile

It's time-intensive precisely because the hard parts can't be automated. But the output is a properly structured .tex file that compiles cleanly - not a pile of image fallbacks with broken formatting.

Why Word to LaTeX Conversion Breaks (And What to Do About It)

Saurabh Shah — Sat, 21 Mar 2026 10:21:29 +0000

A technical look at why automated tools fail on academic papers — and what actually works.

If you've ever tried to convert a Word document to LaTeX using Pandoc or an online converter, you already know how this ends. The text comes through fine. Everything else doesn't.

We've handled hundreds of Word to LaTeX conversion projects at The LaTeX Lab — IEEE papers, Springer submissions, Elsevier articles, conference proceedings. The same failure modes show up every time. Here's what's actually going wrong under the hood.

Why Pandoc Fails on Academic Papers

Pandoc is genuinely impressive for what it does. But the way it handles equations reveals a core limitation: it converts Word's OOXML math format to LaTeX by pattern-matching, not by understanding mathematical structure.

The result is that simple inline equations often convert correctly. Anything structurally complex — nested fractions, matrix environments, multi-line align blocks, custom operators — gets either mangled or exported as an image fallback. You end up with this:

% What Pandoc gives you for a matrix equation:
\includegraphics{eq_img_01.png}

% What it should be:
\begin{equation}
\mathbf{A} = \begin{bmatrix} a_{11} & a_{12} \\ a_{21} & a_{22} \end{bmatrix}
\end{equation}

Image fallbacks technically compile, but they fail journal submission checks — most publishers require equations in proper math mode, not embedded images.

The Table Problem Is Worse

Word tables store alignment, merging, and cell padding in a proprietary format that has no clean LaTeX equivalent. Pandoc's output for a moderately complex table typically looks like this:

% Pandoc table output — broken alignment:
\begin{longtable}[]{@{}lll@{}}
Name & & Value \\ % ← empty cell where merged cell was
\end{longtable}

A properly formatted academic table should use booktabs with \toprule, \midrule, and \bottomrule. Pandoc doesn't output booktabs by default, and even with the --no-highlight flag it doesn't reconstruct merged cells correctly.

Bibliography: The Silent Failure

This is the one that costs the most time. Pandoc exports citations as plain text — it doesn't produce a .bib file. If your Word document uses Zotero or Mendeley for citations, those come through as formatted strings, not structured BibTeX entries.

Which means you're left manually recreating every entry:

% What you need:
@article{Smith2023,
author = {Smith, John and Doe, Jane},
title = {A Study of X},
journal = {Nature},
year = {2023},
volume = {12},
pages = {45--67},
doi = {10.1038/s41586-023-00001-x}
}

% What Pandoc gives you in the .tex file:
Smith, J., & Doe, J. (2023). A Study of X. \textit{Nature}, 12, 45–67.

For a paper with 30–50 references, rebuilding the bibliography from scratch is a 2–3 hour job on its own.

What a Clean Word to LaTeX Conversion Actually Looks Like

For a properly converted academic paper, you need:

Every equation typeset in LaTeX math mode — $...$ for inline, \begin{equation} or align for display
Tables rebuilt in booktabs format with correct column alignment
Bibliography as a complete .bib file with proper entry types (@article, @inproceedings, @book) and all required fields
Journal template applied and verified — IEEEtran, acmart, elsarticle, llncs, etc.
The whole thing compiled and tested before you touch it

Automated tools handle the first bullet partially on a good day. They don't touch the rest.

The Actual Time Cost of DIY Conversion

Here's what a realistic Word to LaTeX conversion looks like if you do it yourself on a 15-page IEEE paper with 25 equations, 4 tables, and 30 citations:

Run Pandoc, fix encoding issues - 30 min
Re-typeset broken equations by hand - 3–4 hrs
Rebuild tables in booktabs - 1–2 hrs
Rebuild bibliography as BibTeX - 2–3 hrs
Apply IEEEtran template, fix conflicts - 1–2 hrs
Debug compilation errors - 1–2 hrs
Total - 8–14 hrs

That's not an edge case — that's a typical paper. For researchers on submission deadlines, it's a significant hidden cost.

DEV Community: Saurabh Shah

LaTeX for Academic Publishing: What Journals Actually Require (And What's Optional)

Which Publishers Actually Require LaTeX

The Three Most Common Journal Template Mistakes

1. Loading packages the class file already loads

2. Wrong citation style for the journal

3. Two-column float handling

What "Template Compliance" Actually Means in Practice

arXiv-Specific Notes

Further Reading

The LaTeX Compilation Errors That Waste the Most Time (And How to Fix Them Fast)

1. `! Undefined control sequence`

2. `! Missing $ inserted`

3. Package Conflicts (`Option clash for package X`)

4. `! File 'X.sty' not found`

5. `Runaway argument` / `Paragraph ended before \X was complete`

6. BibTeX Errors That Don't Show Up Until You've Fixed Everything Else

7. `Overfull \hbox` Warnings (The One You Should Actually Fix)

The Fastest Debug Workflow

Further Reading

PDF to LaTeX Conversion: Why It's Hard and What Actually Works

Why PDFs Are a Poor Source for LaTeX Reconstruction

What Automated PDF to LaTeX Tools Actually Produce

Tables: Even Harder Than Equations

The Text Layer Problem on Scanned PDFs

What the Right Approach Looks Like

Further Reading

Why Word to LaTeX Conversion Breaks (And What to Do About It)

Why Pandoc Fails on Academic Papers

The Table Problem Is Worse

Bibliography: The Silent Failure

What a Clean Word to LaTeX Conversion Actually Looks Like

The Actual Time Cost of DIY Conversion

DEV Community: Saurabh Shah

LaTeX for Academic Publishing: What Journals Actually Require (And What's Optional)

Which Publishers Actually Require LaTeX

The Three Most Common Journal Template Mistakes

1. Loading packages the class file already loads

2. Wrong citation style for the journal

3. Two-column float handling

What "Template Compliance" Actually Means in Practice

arXiv-Specific Notes

Further Reading

The LaTeX Compilation Errors That Waste the Most Time (And How to Fix Them Fast)

1. ! Undefined control sequence

2. ! Missing $ inserted

3. Package Conflicts (Option clash for package X)

4. ! File 'X.sty' not found

5. Runaway argument / Paragraph ended before \X was complete

6. BibTeX Errors That Don't Show Up Until You've Fixed Everything Else

7. Overfull \hbox Warnings (The One You Should Actually Fix)

The Fastest Debug Workflow

Further Reading

PDF to LaTeX Conversion: Why It's Hard and What Actually Works

Why PDFs Are a Poor Source for LaTeX Reconstruction

What Automated PDF to LaTeX Tools Actually Produce

Tables: Even Harder Than Equations

The Text Layer Problem on Scanned PDFs

What the Right Approach Looks Like

Further Reading

Why Word to LaTeX Conversion Breaks (And What to Do About It)

Why Pandoc Fails on Academic Papers

The Table Problem Is Worse

Bibliography: The Silent Failure

What a Clean Word to LaTeX Conversion Actually Looks Like

The Actual Time Cost of DIY Conversion

1. `! Undefined control sequence`

2. `! Missing $ inserted`

3. Package Conflicts (`Option clash for package X`)

4. `! File 'X.sty' not found`

5. `Runaway argument` / `Paragraph ended before \X was complete`

7. `Overfull \hbox` Warnings (The One You Should Actually Fix)