This is probably a FAQ but didn’t see it discussed:
Could anyone fill us in on the major differences / pros-and-cons of fastdoc vs. jupyter book? I am someone who is “crazy enough to think writing a book is a good idea.”
I came to the forum after seeing Jeremy’s tweet today, and searched the forum to find a thread on the relative merits of fastdoc and jupyter book, and/or why jupyter book wasn’t meeting the needs of Jeremy & Sylvain --presumably because (I’m guessing)…
so many custom aspects of nbdev
and/or perhaps that they just wanted/needed to have more control over the project to do what they wanted.
jupyter book layout may not look like what they wanted
much of jupyter book involves rendering an online format, which J&S didn’t want or need for Fasbook.
Jupyter Book didn’t exist in a stable/usable form when we starting writing our book, so we didn’t look at it closely. Also, we were already familiar with the nbdev approach and it worked just the way we wanted it to.
You should try both and see which meets your needs better.
One other note: I’ve also recently become aware of Manifold, which is primarily a book-publishing platform but the build process can use a Markdown-based assembly similar to Jupyter Book, fastpages & such systems.
Hi @drscotthawley, my co-authors and I started with jupyterbook, but recently switched to fastdoc because of the asciidoc support that some publishers require.
Overall jupyterbook is very well designed and would be great for publishing lecture material or a textbook you want to publish for free on e.g. the arXiv. It also comes with a lot nice features like BibTex support, launch buttons for Colab, Binder etc, and handy navigation menus to quickly switch between chapters and subsections.
I have only just started using fastdoc, but one obvious advantage is that I can use nbdev and don’t have to learn yet another syntax. As a new project, the docs for fastdoc are somewhat sparse, but I hope to improve that as we write the book
