May 30, 2020By Joe Lafiosca← Back to Blog

The Terrible Homepage


My name is Joe Lafiosca. I used to develop many sites and domains back when that was an unusual and difficult thing to do. Then because I was preoccupied with other things such as life and work, I allowed nearly all of them to languish and decay. I decided it was finally time to create a new one for myself.

Today we are in another time. There are so many different options out there for building and hosting web applications that it can cause a sort of decision paralysis. Beyond that, our standards for what is a reasonable level of effort in self-hosting have changed over time. Many choose to use canned website service offerings, but I wanted something that would:

  1. allow me to use and integrate with my existing provider, AWS
  2. not require a new persistent server (VM, container, database, etc)
  3. not have any significant cost

With these goals in mind, I figured I'd like to use a static site generator which could be deployed to S3. The last time I examined these technologies, in January 2019, I identified Gatsby as worth exploring, largely because it was based on React, with which I was already comfortable developing. At that time, my goal was to migrate a multi-author Ghost blog running persistently on ECS and RDS into a serverless configuration. I quickly decided that I needed some form of CMS to manage that, and Netlify CMS seemed it might fit the bill. I wanted an S3-based backend (as opposed to git) for the Netlify CMS content, and that did not exist, so I started working on one. Unfortunately, the backend storage API was not yet fully baked, my progress was slow, and my attention ultimately drifted to other projects.

Fast forward to the present, when rather than dive back into Netlify CMS, I thought I'd see what other options were available. TinaCMS looked promising, so I decided to try out the Tina Grande Gatsby starter. Already in the course of writing this article, I've somehow managed to crash it numerous times, but there seems to be a lot of potential here. And I figure even if I ultimately move to a different solution, having my articles as markdown files in a repository makes for a pretty flexible resource.

Now I don't have the patience nor interest in digging deeply into the Gatsby side of this project to modify its functionality, and I'm not certain I ever will. To be frank, without intending to be disparaging, I am not a fan of many of Gatsby's style decisions, and I am disappointed with its TypeScript support. My goal today was to start up and deploy a solution that would be "good enough" for my homepage and blog until such time may come that I have the energy to build something more novel, rather than allow my personal domain to continue to sit uninhabited. And I am hopeful that we have reached that goal.