I’m very puzzled. You are operating from the assumption that 34 kB would take 50 seconds to download on a 1mbs connection? There was clearly something else at play, maybe related to the fact that the people who built his react codebase had never programmed before.
I'm not operating from any assumptions, I am just stating observations.
The observations were:
- ~6MB JS file produced by `npm run build` compared to ~400KB page weight for all items in my vanilla app
- Page load times from that Greek WiFi was ~4 secs for the vanilla app and ~50 secs for the React app, which told me the throughput was on the order of 1 Mbit/s. The React app didn't load many things (about 3 resources, if I remember correctly) while my app had more to do. This means network or server delay will have mainly affected load time of my app.
Of course, this whole performance/page weight thing wasn't the main point of the article (for those who happened to read it), nor did it very much affect my decision to go vanilla. I am surprised it became such a big issue here, but I've stated that I too think it was due to inexperienced developers. I guess my mistake was thinking that my lack of front-end experience would make it a viable comparison - I'm probably a lot more performance-conscious than most, and especially compared to someone who is just learning to program.
