Sep 28, 2025, 8:20 PM -04:00

Implementing a Reading List in Org Mode

I joined Goodreads in 2012, at roughly the same time I joined Twitter, in what might be called a more innocent age of social networking. The idea of a social network for book readers appealed to me at the time.

Fast forward to 2025, and the world looks different. Twitter/X has turned into a hellscape run by a super villain, and Goodreads, though not quite a hellscape, is slow, dated and run by a company with dubious ethics.

I ended up deactivating my Twitter account - not that difficult given that I never really gained a following. As it turns out, my use of Goodreads as a social network never really took off either - I don't think my friend list ever exceeded 20 people. I used it much more as a log of my reading habits, a record of every book I ever read or wanted to read. That made it easier to think about finding a replacement.

Alternatives to Goodreads

There are several viable alternatives to Goodreads particularly if, like me, you aren't that invested in the social aspect of maintaining a reading list. In 2025, the contenders seem to include, but are not necessarily limited to:

BookWyrm, in particular, is interesting as it's part of the Fediverse and uses the ActivityPub protocol. It's therefore truly decentralized service, like Mastodon.

Here's the thing, though. Ever since I began participating in the IndieWeb back in 2019, I've been been trying to live up to its principles - in particular the part about owning your data. And while all these services may have the best of intentions, there's no predicting the future. Any one of them might go out of business, or might simply evolve to the point where you want to leave - like Twitter did.

So what can you do?

Settling on Emacs and Org Mode

If you read my blog, you might know that I'm a fan of keeping my data in plain text. You might also know that I'm an Emacs enthusiast who cobbled together a vaguely GTD like system using Org Mode. Given that a reading list is, fundamentally, nothing more than a specialized TODO list, it makes sense to try and do something similar here.

Please note that, as usual, this will not be an Emacs or Org Mode tutorial. If you want to learn more about these tools, please consult the fine documentation.

So how do we go about doing this? Plain vanilla Org mode itself gets us about 80% of the way there. Here's a small example file:

#+TODO: TOREAD READING | READ DNF

* READ The Neverending Story  :fantasy:youngadult:translation:
CLOSED: [2025-03-06]
:PROPERTIES:
:Author:   Michael Ende, Ralph Manheim
:Added:    [2015-07-24]
:Rating:   :star::star::star:
:END:

* TOREAD The Language Instinct: How the Mind Creates Language     :linguistics:
:PROPERTIES:
:Author:   Steven Pinker
:Added:    [2025-01-12]
:END:

There's nothing fancy here. The first line defines the TODO states that apply to this file. Since this file is a reading list, we have two actionable states: TOREAD, READING and two terminal or DONE states: READ and DNF (for Did Not Finish). Org mode automatically puts in a CLOSED line (as seen above) when the state of a headline moves from an actionable one to a terminal one. You can use this to record when you finished (or did not finish) a book.

Each title is entered as a headline with a TODO state. Extra information like Author and Added date is recorded via org mode's properties feature. I've also leveraged org mode's built in tagging system to define a basic taxonomy for the reading list.

To display your books using org-agenda, you can add this section in your org-agenda-custom-commands block:

("r" "All Books"
 ((todo "READING")
 (todo "TOREAD")
 (todo "READ")
 (todo "DNF"))
 ((org-agenda-files (file-expand-wildcards "~/org-roam-garden/books.org"))
 (org-overriding-columns-format "%TODO %50ITEM(TITLE) %30AUTHOR %30RATING")))

Now you can type "r" when calling up an org agenda and you'll get a basic, compact reading list. Even better, you can leverage org mode's agenda column view by typing C-c C-x C-c to get a nice tabular format.

Leveraging org-books

You can type out the books in your reading list manually, but since this is Emacs, it would be weird if there weren't a way to automate the process.

Help exists in the form of a Emacs package called org-books which advertises itself simply as "Reading list management using org-mode". Mostly, that means it provides a collection of elisp functions to quickly add a book to your reading list via a URL (org-books-add-url) or ISBN (org-books-add-isbn).

The org-books-add-url function, in particular, claims to support Amazon, Goodreads and Open Library as sources. What this means is that you can interactively call org-books-add-url and paste in, for example, an Amazon book URL when prompted. Press enter and voila - a new entry will be added to your book file. It looks like this:

* TOREAD Piranesi                                                       :fantasy:
:PROPERTIES:
:AUTHOR:   Susanna Clarke 
:ADDED:    [2015-07-31]
:AMAZON:   https://www.amazon.ca/Piranesi-Susanna-Clarke/dp/1635577802
:END:

Formatting with an Eye Towards Export

To be honest, I found org-books a bit rough around the edges. While Amazon and Goodreads URLs seemed to work, Open Library URLs would only work with the book's JSON API endpoint, not the actual web page, which I found annoying.

In addition, I wasn't totally happy with the way the package formatted the book entries. Informing my opinion was my desire to be able to easily export my reading to HTML, for publication on my garden.

Org mode isn't primarily a CMS and while it does support HTML export out of the box, it doesn't really come with what I would describe as an easy to use templating system. What this means in practice is that, without some fairly major configuration, org-mode HTML export will result in HTML that more or less mirrors the original content in terms of structure and semantics.

What I'm trying to say is that if you want your exported HTML to be structured a certain way, it's often easiest to arrange for your org-mode source to be structured in a similar way.

Related to this is the fact that org-books, seemingly as a matter of policy, formats all of its book entries exclusively as "headlines + properties", presumably reserving the content area for extra comments about work itself.

This can be a problem when it comes time to export the entry to HTML, because properties are treated as special in org-mode's export system - they are rendered as plain text by default, not HTML. So that URL in the AMAZON property you see above? Because it's a property, it's rendered in HTML as a plain, unclickable URL. After some amount of research, I came to the conclusion that the easiest way to get an actual link in the output was to embed the Amazon URL as a real org-mode link, in the content of the entry, not the properties - since proper org-mode links do get turned into HTML links during export, as you might expect.

I also thought it would be nice to have display cover image for each book entry, particularly in the exported HTML, and Open Library does have a public covers API that I suspected I could leverage. Org-books, so far as I could tell, doesn't insert any cover images at all. In any case, the right place for it, assuming we wanted it to appear in the exported HTML, would be as an image in the content, not as a property.

All this to say, org-books wasn't totally cutting it. So what to do?

Hacking org-books

In the end I decided to fork org-books and make some changes:

I capitalized the data properties, but I didn't make them all uppercase like they were originally. The property names appear in the HTML as is so I wanted something less "in-your-face".
I embedded the source URLs as proper links in the entry's content, rather than as properties.
When adding a book via Open Library, I also inserted a link in the entry's content to the cover via their covers API.
I added a function, org-books-add-olb, that accepts an Open Library web page URL as a data source, in a similar manner to Amazon or Goodreads.

I made sure that the entries were formatted in a way that made it easier to later produce the layout I wanted via CSS. In the end, a typical entry looked like this:

* READ Redshirts                   :scifi:
CLOSED: [2025-01-28]
:PROPERTIES:
:Author:   John Scalzi
:Added:    [2015-07-31]
:ISBN:     9780765316998
:Rating:   :star::star::star::star:
:END:

#+BEGIN_aside
#+ATTR_HTML: :loading lazy
[[https://covers.openlibrary.org/b/ISBN/9780765316998-M.jpg]]

[[https://openlibrary.org/books/OL25253043M/Redshirts][Open Library]] \\
[[https://openlibrary.org/api/books?bibkeys=ISBN:9780765316998&jscmd=data&format=json][Open Library Data]]
#+END_aside

I expected this book to be funny, and it was.  What I was *not* expecting
was that this book would make me emotional.  I nearly cried.

The content has a preamble that I wrapped in a #+BEGIN_aside/#+END_aside pair; when exported to HTML, this results in the enclosed content being wrapped in an aside tag that can be easily styled by CSS.

The formatting of the cover image is worth noting. It looks like this:

#+ATTR_HTML: :loading lazy
[[https://covers.openlibrary.org/b/ISBN/9780765316998-M.jpg]]

This results in a loading="lazy" attribute added to the cover img tag in the exported HTML:

<img src="https://covers.openlibrary.org/b/ISBN/9780765316998-M.jpg" alt="9780765316998-M.jpg" loading="lazy">

The effect is to only load the image when it's scrolled into view. Given that I have hundreds of books in my reading list, the exported page loads much faster as a result.

Example and Final Thoughts

I publish my reading list to my garden. You can peruse it there.

Is this thing a Goodreads killer? Will it replace a real book tracking app for the average person? Good God, no. I'm the first to admit the whole process is clunky as hell.

But apps comes with trade offs, as any IndeWeb enthusiast is aware, and one of the things you give up when you use one is full control over your data. The one big advantage of all this trouble is that my reading list is mine and mine alone. I consider that important.

Important enough to always resist the siren son of a dedicated service for years to come? Maybe not - that's why the IndeWeb has a prescribed way for sharing your data with the rest of the world: POSSE. This is the practice of publishing to your own site first and then deciding how you tell the world about it. It's something I do on my blog - all posts are automatically shared to Mastodon once published.

To be honest I'm not sure how POSSE would look in this case; I'm not even sure what service I'd share to, or how I'd figure out what to share. Still, it's an interesting idea worth exploring.

In the meantime, this serves my needs, and it was fun to do.