Scott Nicholas writes at the Linux Foundation blog:
A key goal of some open collaboration efforts — whether source code or specification oriented — is to prevent technical ‘drift’ away from a core set of functions or interfaces. Projects seek a means to communicate — and know — that if a downstream product or open source project is held out as compatible with the project’s deliverable, that product or component is, in fact, compatible. Such compatibility strengthens the ecosystem by providing end-users with confidence that data and solutions from one environment can work in another conformant environment with minimal friction. It also provides product and solution providers a stable set of known interfaces they can depend on for their commercially supported offerings.
A trademark conformance program, which is one supporting program that the LF offers its projects, can be used to encourage conformance with the project’s code base or interfaces. Anyone can use the open source project code however they want — subject to the applicable open source license — but if a downstream solution wants to describe itself as conformant using the project’s conformance trademark, it must meet the project’s definition of “conformant.” Some communities choose to use words other than “conformant” including “certified”, “ready”, or “powered by” in association with commercial uses of the open source codebase. This is the approach that some Linux Foundation projects take to maintain compatibility and reduce fragmentation of code and interfaces.