I’m rather smitten with the *idea* of dev containers and, to a degree, the reality delivers. Reproducible environments. Familiar primitives. Editors with decent adoption offer a good experience. Most engineers understand containers well enough to tolerate the rough edges.
But I can’t shake the feeling there are better ways to reach the same goal.
The usual answer is “Nix”, and that’s probably right. Yet the rise of Dev Containers feels less like a breakthrough and more like a side-effect of the container monoculture. They work because containers are everywhere, not because they’re especially well-suited to dev environments. If there’s a real successor, it likely abstracts Nix *away*, integrates cleanly with editors, and doesn’t require heavyweight plugins or full VM-scale artefacts to edit a file.
What’s wrong with dev containers? They’re slow, expensive, and bloated relative to the problem they’re solving. I’m still invested, and I’ll keep using them. But it does feel like, as an industry, we’ve settled for the first thing that worked rather than the best thing we could build.
#DevContainers #docker #vscode #zed #vim #jetbrains #python