RFC 136: Looking for feedback on the scope

Hi everyone,

RFC 136 (“A plan to stabilize Flakes and the new CLI incrementally”) has been going through a number of iterations. In particular, it saw its scope expanded a lot, from a simple roadmap towards stabilizing the new CLI to a large set of general principles around the development of Nix.

The reasoning behind that is that it was sometime unclear where exactly the “CLI, then flakes” plan was coming from. That lead to us writing a bunch more background material on layering/modularity, and stabilizing things in general. The idea was that the broader material could both provide an “origin story” for the plan, and (hopefully) ensure we don’t get into this situation again. The counterargument is that it makes this RFC too big and unfocused and that we should rather focus on solving the problem at hand rather than trying to prematurely abstract over it.

This now requires some more feedback, and in particular answering the question: Is the expanded scope of this RFC good, or should the new material be left for the future?

Please sound off in the comments, letting us know whether you think we should keep or cut this new material! That will inform what we do next, and then we can hopefully wrap this thing up.

I personally have my own very strong opinion about that, but I’ll refrain from mentioning it in this post


To more effectively produce mind share on the architectural considerations that support the origin story, I’d enjoy to see more use of architecture diagramming in all communications.

I do understand that origin story, but it’s hidden behind a wall of RFCs, comments, text, context and trust in what protagonists stand for that I’ve personally gathered.

If the design language was visual, it potentially just snaps right into the audiences brains.

Overall, my understanding is such that not an argument must be won, but shared context must be built.

When are flakes going to be stabilized looking at this?

Last activity on the PR is more than a month ago. What’s the plan?

I think the scope of the RFC is fine. I understand why it became what it became, but I also feel like it’s important to have the ideas behind it in writing.

However, there’s still TODO items in the RFC, so it doesn’t seem to me like it’s even finished.


OK. It seems:

RFC 136: A plan to stabilize Flakes and the new CLI incrementally - pinged Tom to suggest a meeting for potentially calling FCP

That the RFC may go to the next stage based on the meeting notes from yesterday.

Thanks for the reminder. I’ve fixed the TODO.