Hacker News Reader

Comment by chillpenguin | Hacker News Reader

I only get a new laptop once every several years. Doesn't really seem worth it to me personally. I also sort of like starting fresh in a way. Granted I have my dot-files on github to make that part easier. But I don't mind running the install command for things as they come up.

I'm curious if there are other benefits to this approach though besides just saving time when setting up a new machine. The article mentioned "you end up with (in my opinion) a much more flexible working environment." Any ideas what they might mean?

View on HN

Reproducible dev environment. Easier to reproduce some bugs for fixing. More certainty that it doesn't just work on your laptop because you have an undocumented dependency installed. Easier to test the setup process on a clean machine and vary things about the machine setup. To test what happens if you have Python available system-wide vs. if you don't. More precise development history since you have docker-compose.yml under Git, making "time travel" easier.

There's all kinds of little benefits that don't seem that important until you have use for them. Of course Guix and Nix go closer to being actually reproducible, but Docker is better than nothing.

To add maybe one more point to this - it’s so much easier to run parts of the pipeline for devs who aren’t familiar with the environment. It can also serve as a documentation for what’s required to develop.

I now go a bit further. I used to keep my dot files around as well but last time I decided to go completely fresh and I learned about powerline10k (more performant than powerline9k) and sdkman (an installer for multiple and different sdk flavours) and had a nice evening to boot. If I hadn’t started over I would just have used the old config and not enjoyed the new benefits.

I think I share your line of thinking.

What are the benefits? Are there down sides to being operating in the docker container for everything?

Reply on news.ycombinator.com