This website now feels finished enough to deserve a proper description.

Not finished in the absolute sense, because websites are rarely finished, but finished enough to have a clear internal logic. It is no longer just a personal site assembled page by page. It has become a small publishing system shaped around the kind of work I actually do: papers, projects, public resources, open questions, and a more reflective kind of writing alongside them.

I have decided to call this template Arvel Scholar.

The lineage

The site is built with Hugo and ultimately rests on the PaperMod theme. That is the technical foundation.

But it would be misleading to say that this is simply a PaperMod website. The site also builds on Pascal Michaillat’s Hugo website template, which is itself a highly thoughtful academic reinterpretation of PaperMod. Pascal also wrote about that template here: A Hugo website template for academics.

So the path is straightforward:

  • PaperMod
  • Pascal Michaillat’s academic customization of it
  • my own further customization for a bilingual research website

That last stage is what I am calling Arvel Scholar.

Why I changed it

The main issue was not appearance alone. It was structure.

I did not want one undifferentiated stream of posts. I wanted the site to distinguish more clearly between different kinds of academic work:

Those categories do not behave the same way, and they should not look as if they do. A paper is not a project. A resource page is not a reflective note. A question is not a blog post. The site therefore needed more than a theme; it needed clearer internal rules.

There was also a practical reason for the redesign: before this consolidation, parts of my web presence were spread across separate systems.

  • the main personal site was on WordPress
  • the old Questions site was a full Jekyll site using the Chirpy theme
  • the old Resources site was another full Jekyll site using Just the Docs

So this was not just a theme customization exercise. It was also a migration away from fragmentation.

The main difference from Pascal’s version

The homepage is where the divergence became most visible, and two features matter more than any others there:

  • dark and light mode
  • English and French language support

Those might sound like modest additions, but they fundamentally change the site.

Pascal’s version gave me a strong academic presentation model. What I needed on top of that was a site that could move comfortably between languages and that could feel usable across different reading conditions. That meant the homepage had to become not just elegant, but operational.

Part of that shift is visible directly in the configuration:

defaultContentLanguage: en
defaultContentLanguageInSubdir: false

languages:
  en:
    languageName: English
    contentDir: content/en
  fr:
    languageName: Français
    contentDir: content/fr

params:
  defaultTheme: auto
  disableThemeToggle: false

The meaning of that block is simple: this site is built to move between languages, and it is built to move between visual modes.

The theme toggle is equally small in code and equally important in practice:

document.getElementById("theme-toggle").addEventListener("click", () => {
  if (document.body.className.includes("dark")) {
    document.body.classList.remove("dark");
    localStorage.setItem("pref-theme", "light");
  } else {
    document.body.classList.add("dark");
    localStorage.setItem("pref-theme", "dark");
  }
});

This is not a spectacular technical innovation. But it changes the way the site is inhabited. It makes the homepage less like a static statement and more like a working surface.

The language layer is just as important. In the earlier Jekyll sites, multilingual logic was not part of the overall architecture. Here it is built directly into the site structure:

content/
├── en/
│   ├── papers/
│   ├── projects/
│   ├── resources/
│   ├── questions/
│   └── writing/
└── fr/
    ├── papers/
    ├── projects/
    ├── resources/
    ├── questions/
    └── writing/

That shift is one of the reasons the homepage now feels more integrated than the older sites.

What changed in practice

The homepage now acts less like a generic landing page and more like an orientation page. It introduces the site as a research environment rather than a simple list of recent content.

Beyond that, I spent time making the internal sections more legible.

Papers

The papers section had to remain plain and citation-friendly. I wanted it to read as a research archive rather than as a promotional page.

Projects

The projects section needed stronger internal framing. Projects have their own timeline, status, and purpose, and they should not be visually collapsed into publications.

Resources

The resources section needed cleaner navigation and a more practical layout. These pages often function as documentation, teaching materials, or public references rather than as ordinary posts.

This section is also one of the clearest examples of migration in practice. Before this Hugo site, Resources Wall was its own Jekyll documentation website using Just the Docs. In its old configuration, the theme declaration looked like this:

remote_theme: "just-the-docs/just-the-docs"
plugins:
  - jekyll-seo-tag
  - jekyll-github-metadata
  - jekyll-include-cache
  - jekyll-sitemap
  - jekyll-remote-theme

That setup worked well for documentation, but it also kept the resources in a separate publishing universe. Migrating them into Hugo meant they could sit inside the same research website while still keeping a clearer resource-specific presentation.

Questions

The questions section is one of the most specific parts of the site. It is closer to a structured intellectual notebook than to a conventional category archive.

This part of the site also has its own earlier history. Before being folded into the current structure, Questions was a separate Jekyll site built with Chirpy. Its theme configuration was explicit:

theme: jekyll-theme-chirpy
title: Questions
tagline: A collection of the questions that keep me awake
url: "https://questions.mkante.ml"

That earlier site already had a strong identity, but it remained disconnected from the rest of my publishing. Rebuilding Questions inside Hugo was less about preserving the old theme and more about preserving the intellectual function of the section while integrating it into a larger whole.

Some of that differentiation appears directly in the templates:

{{ $isQuestionsPage := eq .Section "questions" }}
{{ $isProjectsPage := eq .Section "projects" }}

{{ if $isQuestionsPage }}
  <div class="questions-shell">
    ...
  </div>
{{ else if $isProjectsPage }}
  <div class="project-single-meta">
    ...
  </div>
{{ end }}

The point is not the snippet itself. The point is that the site now accepts that different forms of academic work need different display logic.

Migration, not just redesign

Another important part of this story is that the present site was not created only by hand-copying content from one place to another. The main site itself came from a WordPress to Hugo migration workflow.

Hugo’s official migration page is here: Migrate to Hugo. It points to tools for moving content from several systems, including WordPress and Jekyll. In the WordPress case, it specifically lists wordpress-to-hugo-exporter, which is the path I had in mind when working through the exported material.

At a very rough level, one part of the change looked like this:

# Jekyll
bundle exec jekyll serve

versus:

# Hugo
hugo
hugo server

That comparison is simplistic, but it captures something real. The move to Hugo was also a move toward a faster, cleaner, and more unified build process.

Why the name Arvel Scholar

I considered other names, including Reflective Scholar Mode, but that one sounds more like an interface setting than a template identity.

Arvel Scholar feels more appropriate. It sounds like a template that could stand on its own: a derivative academic website system, rooted in an existing theme tradition but adapted for a more personal and multilingual use.

It is not meant to replace PaperMod, and it certainly does not erase Pascal Michaillat’s contribution. On the contrary, the point of this work is precisely that careful adaptation can be more productive than starting from nothing.

What this site now represents

At this point, I no longer think of the site as “my PaperMod site.”

It is better described as a bilingual research website built from:

That final layer is where the site becomes mine.

And that, for now, is what Arvel Scholar names.