Marcel Kapfer

Update on my Org-roam web viewer


453 words, ~4min reading time

100DaysToOffload emacs orgmode dev

About one and a half weeks ago I announced that I'm going to start a new (code) project for viewing my Org-roam files over the web (especially on mobile devices). Since then surprisingly much happened and so I want to give a short status update.

First of all I chose a working title for the software: brainengine. However, I'm not quite sure why. Especially the engine part since it doesn't power anything. Though the brain makes sense since some people (myself included) refer to Org-roam and similar software as their second brain. But the name may change.

Concerning the perhaps more interesting stuff: I made at lot more progress during last and this week that I expected. To be precise: as of know nearly ever core functionality works. Although I went a different route in the end.

My initial goal was to create a “classic” client-server application using Go for the backend and Alpine.js with tailwindcss for the frontend. That's why I started writing an API with Go in the first place. Parallel (and initially completely unrelated) I worked through the new Templating tutorial of Learn Go with tests by Chris James and started thinking if I maybe should use this for a start. And so I started playing around wit Go's HTML templating and as of this evening I not only have a API (that I don't use and perhaps not even need) but also an HTML representation to my Org-roam files. Both have the same feature set:

And I did neither use Alpine.js nor tailwindcss in the end but only Go and Bulma for some basic styling.

Concerning the underlying concept nothing has changed compared to the original idea. Every single function (except reading a file) uses some Linux/UNIX command line tool in the end. bash + ls (with globbing) is used for getting a list of all Org-roam file (this will be the first thing for which I want write a Go-only solution) and grep for searching through the files and the getting the title of a node as well. Sadly the code is currently still that crappy that I'm not willing to share it yet. Perhaps I will only open-source the final application since the current implementation is only a proof-of-concept (I wanted to see where this approach might fail and I have to say: nowhere! Even the load times are by far fast enough for now).

Of course there are currently still some things that do not work or are not yet implemented:

Besides these four points there are certainly many other things that I need to improve. I'll write again a blog post once there are new things to tell.

Day 7 of the #100DaysToOffload challenge.

I would like to hear what you think about this post. Feel free to write me a mail!

Reply by mail