Companion RSS reader for my blog, created using TinyTinyRSS and Heroku

I started blogging in 1999, well before the word ‘blog’ was commonly known. I remember saying ‘blog’ to people and having them respond “What?”

One of the things I learned from those early experiences was that writing a good blog came from reading widely. Inspired by Dave Winer’s River of News, I wanted to create my own river of sources, just like I used to have.

I had hoped to make mine public, the way Dave does. I haven’t quite got there yet. The codebase I’m using as a starting point, TinyTinyRSS (ttrss) reminds me a lot of FeedDemon, my longtime favorite desktop RSS app – it’s just that it’s hosted on the web. I have to log into it to see anything. The use case is one user reading the most recent updates from their favorite news sources privately. Still, I did manage to get a web-based RSS reader up and running by forking, cloning, and setting up the codebase myself, and I thought I’d document how I did it.

For those of you who have not heard of RSS, it is a file format that has a machine-readable version of a blog or other online news source’s stories. The feed for this blog looks like this:

screenshot of RSS feed for my blog

Now, that’s not something you’d want to read in your web browser – it’s , but it’s a structured version of the content of this blog that an application can read.

I chose TinyTinyRSS, (ttrss), an open-source web-based RSS reader. Since it’s web-based, I needed somewhere where the app and my reading list of sites to be. I was hoping not to shell out any additional $$ for hosting, so I was thrilled to find ttrss-heroku, which allowed me to set up the application on Heroku, an application hosting service with a generous free tier for hobby users and small projects.

After putting it on Heroku, it was, ta-da! live on the web. I began adding a few feeds. One thing I have yet to tackle is automating the update of those feeds; ttrss needs to go read the feeds I subscribe to and fetch any new items for me to read. I’m not totally certain that the automated job scheduler I set up on Heroku is actually working, but I do know how to tell ttrss to update the feeds manually from the command line.

Honestly, I find even that kind of satisfying. Both this blog and my ttrss setup are highly manual. If they were a vehicle, they’d be a dune buggy. I write my entries in a text editor, and commit and publish them from the command line. I enjoy it, and I think I’ll enjoy going back to even more of my old-school ways of blogging and reading, just as I did so many years ago. It’s good to escape from the massive platforms like Facebook and Twitter, and have a simple and independent existence, thinking and writing out here on the open web. I feel liberated by it. It all still works, and it’s all still good.

Update: the scheduled update task I created on the Heroku server is updating my subscriptions. That’s good too ;-)

Update update: When I tried to reload my RSS reader, it hung at ‘Loading, please wait’ until I ran heroku stop <dyno-name> and then heroku restart. Fascinating.