Migrating this website to org-roam

In “Beginning Blogging with org-publish” I had stated that “I want this website to be both a blog of what I’ve done and a sort of living knowledge base with my most current configurations.” We’ll three years later, having failed to update this site with any new content, it’s time for some reflection. In that time I started actually doing “real” academic research, and it has felt like a perpetual struggle to be organized. Hopefully I actually write up what my systems for managing this chaos seem to be converging to, but here I just wanted to note what may have gone wrong with my initial ideas for this space.

Blog-style posts encourage dead content

Most of what I record in all my org-mode files is not “one-shot” content like a blog-style post where once it’s written, it stays relatively unmodified. I’m constantly adding to and revising these files, and if I’m concerned about preserving some sort of “sanctity” of the content as it was originally published, then I feel reluctant to touch those files and that information, even if it is just for minor corrections and modification. In the process of writing, even technical writing, one needs to acknowledge how the process feels. So I have to realize that small changes in the look and organization of my org-mode files can actually have profound effects on my willingness to engage with them. Concretely, I think two decisions early on made created a bit of an emotional barrier to touching the org-mode files that create this site:

Each file was named something like 2019-02-04-beginning-blogging.org. I think having the date in the filename made me hesitate to touch the file, as I felt like that content was simultaneously unmodifiable and dated.
There was nowhere to record a last modified date. I had thought this would be unnecessary since I keep all my files in git and the git history is a way more powerful record of the history of changes. However, I think going forward, having a modified date will help me feel like it’s okay to revisit these files and make minor corrections or reformat them.

Tagging content causes creation paralysis

Introducing tags for categorizing the files here was a mistake. The rise of roam-style note taking has made me realize that hierarchical and/or categorical note taking can be particularly harmful for me as it adds just enough overhead to starting to write something that I will decide not to engage when I don’t a priori know where the thoughts already belong. Since I had dedicated some non-insignificant amount of time getting my org-publish config to support creating lists based on tags, I felt like this site was stuck with that way thinking. Even the blog Funnily enough I think the fact that that content spans two posts illustrates the previous point about dead content:

A practical issue with the solution I came up with there is that while this worked for publishing, it didn’t help much with me quickly being able to see what that set of posts would be in org-mode itself since I would just have some snippet in the file that looked like

The following blog posts pertain to the evolution of this website and blog:
#+begin_src elisp :results drawer :exports results :eval yes
(my-blog-posts-to-list (my-blog-filenames-for-tag "blog"))
#+end_src

Backlinks:

My current workflows