It only feels odd because that is genuinely an incredibly effective means of generation, and we found it very early on because steam is so fundamental. Nothing wrong with sticking to the best method ever discovered.
Copyparty was mentioned here just the other day, and I started using it this weekend. There were comments about security risks though, based on being a small project with a LOT of integrations. Not sure how safe I feel.
Docker path mapping is needed to let Copyparty show the files I want to access. I run my containers on a server next to my NAS that hosts my files; that's a little complicated.
I work in the software industry and I have a guess regarding what the might do to "fix" the problem.
First, we look for the cause, but in this case it is external: we can't prevent solar flares. So we will turn to mitigation instead:
Data gets flaky and erratic unter radiation, so what we would do is to double- and triple-check the data bits. By adding more levels of data correction, more bits can be wrong and we can still figure what it was supposed to be.
Adding more corrections means more overhead and slower performance, but it can still be made to work within the given constraints of real-time processing. They will need to find a balance between hardening and usefulness.
It took me ages to even get it running at all. Maybe it's more complicated as a Docker container, I don't know, but it needs to run alongside manyther selfhosted services.
And once I finally got it working, it was slow as molasses even on LAN, nevermind access from outside. I never started using it.
Imagine being so much snowflake that a simple font upsets you. Words fail.