I’ve tried using personal wikis in the past, in the form of a TiddlyWiki, but it felt weird to use a program separate from Emacs for that task. I kept some notes files in org-mode, used Deft to sift through them etc. But I can’t really say I felt confident about this system, it just seemed like a lot of information which I couldn’t really keep track of and instead of consulting my reference material I usually searched for the information again.

The idea of a personal wiki is to use it as reference material when you learn something new¹, or to connect your thoughts: a way to unload your brain. I guess it has some similarities to a journal, but less focused on chronological order and personal experience?

Concept mapping was unknown to me before I heard of a piece of software called The Brain², which was mentioned in this reddit post on knowledge management in Emacs. A concept map is similar to a mind map, but doesn’t focus on a central topic as much as a mind map usually does. Instead it is a way to connect different concepts with one another, and the same concept can be linked to from different concepts, creating a network of linked concepts. This is similar to a wiki, but in a visual form³.

I find that a problem with mind maps is that they get cluttered fast; it can be hard to get an overview due to information overload. The Brain solves this beautifully by centering on a node (a thought in brain lingo) and only letting you see its parents, its siblings, its children and related nodes. I was intrigued by this piece of software, but as with TiddlyWiki I thought that it would be in line with my workflow if it was integrated with Emacs. So I started working on a project I call org-brain.

org-brain is pretty similar to The Brain, but has less features and a text based interface, it’s Emacs, after all! It works by having a directory full of org-mode files, which are linked to one another using brain: links. I call these files org-brain entries. You can visualize and navigate the network of entries using M-x org-brain-visualize, and add new children/parents from there⁴. Apart from children, parents and siblings, the concept map buffer also show headlines in the entry file, and external links. These external links are called resources in org-brain lingo, and is a way to connect webpages, files, etc to your entries. New resources can be added from the concept map buffer too; the idea being to quickly connect concepts and relevant information to those concepts.

The org-brain project is pretty young, and my main concern is that it might be too slow when having a large ammount of entry files. org-brain doesn’t use a database, instead it parses an entry file to find its children (brain: links to other entries). This is okay, as we only have to parse one file. The problem occurs when we want to find the parents of the entry. org-mode has no way to know which files links to the current one, so we have to parse every org-brain entry file and get every file that links to the current entry. So far I haven’t noticed any speed problems, but time will tell. Perhaps I need to use a database in the future⁵.

You can install org-brain from Melpa, or clone/fork/download it on GitHub. I’ve been enjoying using it so far, but we’ll see if it will stick in the long run. Also I excuse the composition of the screencast, it could be a lot better (and shorter) but hopefully it gives you an idea of how org-brain operates!