Anytime we change the behavior of the experimental feature, the UUID is recycled. This is because behavior changes can break existing code, contrary to CMake’s typical compatibility guarantees.
I’m not aware of any pending changes that would change 4.4’s UUID, but that doesn’t mean we won’t fix bugs (and change the UUID in the process) if they come up.
There is an issue about dropping these magic uuids completely as the feature itself had already grown maturely: https://gitlab.kitware.com/cmake/cmake/-/work_items/28014 - but it’s not clear abour 4.5 time frame (it is scheduled on the October)
the specific values will change over time to reinforce their experimental nature.
As an experimental feature develops, breaking changes are allowed and come with a UUID update. In order for people trying it to distinguish between “this experimental feature doesn’t work” and “this project isn’t aware of a breaking change”, a project must explicitly specify the UUID of which it is aware.