My old snapshot names were variations of “working,” “working2,” and “final.” They stopped being useful as soon as I took a break.
Now the label identifies the base environment and one meaningful change. A small companion note records why I saved it, what is installed, and what I expect to be true after restoration. A snapshot is a saved state, not an explanation of that state.
I also mark disposable branches explicitly. That keeps a quick local experiment from becoming the undocumented starting point for the next one.
Does anyone have a naming convention that is short enough to use consistently but still useful months later?
Separate the human label from a stable identifier. Labels can improve as your understanding changes; the identifier should still point to the same snapshot and its original notes.
My old snapshot names were variations of “working,” “working2,” and “final.” They stopped being useful as soon as I took a break.
Now the label identifies the base environment and one meaningful change. A small companion note records why I saved it, what is installed, and what I expect to be true after restoration. A snapshot is a saved state, not an explanation of that state.
I also mark disposable branches explicitly. That keeps a quick local experiment from becoming the undocumented starting point for the next one.
Does anyone have a naming convention that is short enough to use consistently but still useful months later?