Moving Beyond Zotero: Embracing the Raw Text File

Published: 2026-06-09

It is a well-known fact among those who know me that I am a proponent of minimal computing. Even though my professional life involves AI projects—which are notoriously resource-heavy—I find immense satisfaction in extracting maximum utility from my hardware using the most streamlined and elegant methods possible.

The Academic's Dilemma

For anyone in academia, managing a massive library of sources is a necessity. Whether it's white papers, conference presentations, or journal publications, you need a way to organize the evidence used to construct your arguments.

Usually, this leads people to a bibliography manager. For the vast majority, the "industry standard" is Zotero. While Zotero is an excellent tool, its sheer dominance made me curious about the alternatives.

The Search for the Perfect Tool

I went on a journey to find a different workflow, testing several options:

• Jabref
• Cobib
• Tellico

While each had its merits, a recurring pattern emerged during my migrations. Every time I switched platforms, the process was identical:

1. Export references to BibTeX.
2. Import that BibTeX file into the new software.
3. Repeat when the next tool caught my eye.

The Epiphany: Front-ends vs. The Core

I eventually realized that these applications are essentially just glorified front-ends for BibTeX.

What is BibTeX? It is the file format and processor that comes bundled with almost every $\text{\LaTeX}$ distribution. It handles the labeling of authors, titles, dates, and every other metadata point required for a citation.

In most workflows, BibTeX is treated as a "hidden" intermediary—something you export to just before finalizing a paper, but never something you touch directly. Transitioning from a world of colorful tags and intuitive dropdown menus to a raw text file can feel, frankly, brutal.

Example of a raw BibTeX entry:

@book{abramowitz+stegun, 
  author = "Milton {Abramowitz} and Irene A. {Stegun}", 
  title = "Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables", 
  publisher = "Dover", 
  year = 1964, 
  address = "New York City", 
  edition = "ninth Dover printing, tenth GPO printing" 
}

The Great Migration

In a bid to optimize the ratio of complexity to capability, I decided to commit. I exported my entire library, deleted my reference managers, and dove in.

I expected this to be a short-lived experiment. However, a single .bib file has now served as my primary reference manager for nearly two years.

How it Works in Practice

Managing thousands of lines of text isn't a slog because BibTeX uses field types to categorize data. Beyond the standard author and title fields, the keywords field is the real secret weapon. I use it to track:

• Which project the source belongs to.
• The primary themes/topics.
• Whether the paper has been read.

The Power of the Command Line

Instead of clicking through folders, I use grep to filter my library. For instance, to find all read papers regarding cybersecurity:

grep -E "keywords.*have_read|cyber_security" file.bib

If I need to store deeper insights—like specific methodologies or key findings—I utilize the note field.

The UNIX Philosophy

My goal is always to use tools that adhere to the UNIX-way: small tools that do one thing well and can be piped together.

By treating my references as data, I can stack operations. For example, to count how many times I've cited specific journals:

grep -o journal.* file.bib | uniq -c | sort

Portability and Longevity

Unlike the "pseudo-plain-text" or SQL-based formats used by many managers, BibTeX is actually plain text. This eliminates migration headaches. I can email my references without worrying about compatibility; if a system can open a text file, it can open my library.

Final Thoughts

The learning curve for anything related to $\text{\TeX}$ is notoriously steep. However, the rewards are cumulative. Because $\text{\LaTeX}$ is the lingua franca of academia, mastering these underlying tools doesn't just make me more productive today—it builds a software toolkit that will remain viable throughout my entire career.