reproducible vs reproduced (Re: "Reproducible build" definition in OpenSSF glossary)

[David A. Wheeler]

> On Apr 25, 2025, at 7:45 AM, Holger Levsen <holger at layer-acht.org> wrote:
> 
> hi,
> 
> (just FYI, food for thought, etc)
> 
> a few years^w^w^wmany years ago we discovered that "reproducible builds"
> might not be the best term, but it was already too late to change it.

It was too late to change it years ago. It's even *more* too late to change it now :-).

I think it's a fine term. The other terms you listed *also* have problems.

...

> Rather recently (basically since I created https://reproduce.debian.net/ )
> it occurred to me that "reproducible builds" for another reason is not
> the best term: because noone cares about reproducible builds, because what we
> really care about is whether a build has (or can be) reproduced.

I would seperate those concepts, and there are already terms for two relevant cases:

* Reproducible builds = CAN BE reproduced.
* Verified reproducible builds = HAS BEEN / IS BEING reproduced & verified

I've been using the latter term & I think people immediately grasp what it means.

--- David A. Wheeler