Writing Your Thesis in LaTeX With Your Advisor in the Loop

A thesis is the longest document most people will ever write in LaTeX, and it is also the one with the most feedback loops. Your advisor reads drafts, sends comments, and expects the next version to address them.

Most of the pain comes from how the project is organised and how feedback travels. This post is a setup that keeps a long document manageable and makes your advisor's feedback easy to act on.

Structure the project as one file per chapter

Use a main file that only assembles chapters. Keep each chapter in its own file:

thesis/
├── main.tex
├── macros.tex
├── refs.bib
├── chapters/
│   ├── 01-introduction.tex
│   ├── 02-background.tex
│   ├── 03-method.tex
│   ├── 04-results.tex
│   └── 05-conclusion.tex
└── figures/

% main.tex
\documentclass[12pt]{report}
\input{macros}

\begin{document}
\tableofcontents

\include{chapters/01-introduction}
\include{chapters/02-background}
\include{chapters/03-method}
\include{chapters/04-results}
\include{chapters/05-conclusion}

\bibliographystyle{plain}
\bibliography{refs}
\end{document}

Note that \include starts each chapter on a new page, which is what you want for chapters. Use \input for pieces that should flow inline. Do not include the .tex extension.

Compile one chapter at a time

A 150-page thesis can take a while to build. \includeonly lets you compile only the chapter you are working on, while keeping page numbers and cross-references consistent from the previous full build:

% in the preamble, before \begin{document}
\includeonly{chapters/03-method}

Run a full build first so the auxiliary files exist, then use \includeonly for daily work. Remove it, or list every chapter, for the final build. A common slip is leaving it on and wondering why chapters are missing from the PDF.

Make feedback easy to give

Your advisor's time is the scarcest resource in the project. Reduce the effort of reviewing.

Add line numbers to review drafts. They let your advisor say "line 214" instead of "the third paragraph on page 41":

\usepackage{lineno}
\linenumbers

Turn them off for the final version.

Reuse the per-author comment macros. The same colour-coded commands from my earlier posts work for advisor notes, with one switch to hide them all:

\usepackage{xcolor}
\newif\ifcomments
\commentstrue

\newcommand{\authnote}[3]{\ifcomments{\color{#2}[#1: #3]}\fi}
\newcommand{\adv}[1]{\authnote{Advisor}{red}{#1}}

Send one chapter at a time. A full thesis is too large to review in one sitting. Agree a rhythm, for example one chapter every two weeks, and send a compiled PDF with a short note on what you want checked: the argument, the structure, or the maths.

Snapshot every draft you send

Each time you send a draft to your advisor, save a named snapshot, such as ch3-draft-2026-11-01. When they reply two weeks later, you will know exactly which version they read, and you can compare it with your current text. This matters when you have kept editing while waiting for feedback.

A named version in your editor, a Git tag, or a dated copy of the folder all work. What matters is naming it when you send, not afterwards.

A review cycle that works

1. Draft one chapter, with line numbers on.
2. Snapshot and compile the PDF.
3. Send the PDF with a short note on what to look at.
4. Receive comments, either on the PDF or in the source.
5. Respond by resolving each comment, and keep a short list of what you changed and what you decided not to change, with a reason.
6. Snapshot again and send the next chapter.

Keeping that list of changes is the habit that saves the most time later, because the same comments tend to come back when the full thesis is read together.

Work with your advisor inside the project

You can send PDFs, or you can bring your advisor into the project. If your tool allows it, adding your advisor as a collaborator means they can read the source, leave comments in the text, and see changes as they happen. Some advisors prefer a PDF and a pen, so ask what they prefer before changing how they work.

In LetX, collaborators are unlimited on the free plan, so inviting an advisor and committee members costs nothing extra. The research-team page mentions a supervisor reviewing drafts as part of a typical lab workflow, but it does not describe dedicated advisor tools, so I would rely on the comment macros above and check the current features yourself. I work on LetX, so test it with your own chapters before you move a whole thesis.

Templates, privacy and the final stretch

• University templates. Many universities require a specific class or layout. LetX lists university thesis templates (87 on the templates page, with 40 linked individually, including MIT, Stanford, Oxford, ETH Zürich, NUS, IIT Bombay, BUET and Dhaka University). Always check your graduate school's current guidelines.
• Privacy. Projects in LetX are private by default and the site states documents are not used to train AI models, which matters for an unpublished thesis or one under embargo. Read the security page and decide for yourself.
• Final build. Remove \includeonly, turn comments off with \commentsfalse, remove line numbers, compile from scratch, and read the full PDF for ?? references and overfull boxes.

Checklist

• [ ] One file per chapter, assembled with \include
• [ ] \includeonly for daily work, removed for the final build
• [ ] Line numbers on review drafts
• [ ] A comment macro for your advisor, with a global switch
• [ ] A named snapshot every time you send a draft
• [ ] A running list of comments addressed
• [ ] University template and graduate school rules checked

What is the best advisor feedback habit you have seen? Tell me in the comments. This wraps up the first week of this series, and I will keep posting daily.