truedat101
commented
9 years ago A particularly telling comment was by the last email sent out by someone on the ccnx-users emailing list, stating they'd moved on to a fork of CCNx project because the new 1.0 release is closed source. Whether that is an intentionally political comment or not, the goal of researchers is often to review the how (it is built), and tweak it to produce some new research and prove out new ideas. I realize the old CCNx 0.8.x code is still out there, but doesn't support the new packet format, or the new 1.0 API. Maybe that's another way to get "there" more quickly. Likewise, one could build on top of the CCN-lite release, to speak the protocol with the 1.0 packet format. An interpreted language like Python or JS has some compelling aspects to it that make them easy to pick up, and particularly suited towards easy demonstration, and won't in any way dilute efforts to build industrial grade software that can be licensed by PARC.
To balance the need for licensees vs. getting people using and building CCN 1.0 protocol compliant software, provide a version of the software written in some other language (not Java, not C) that will give developers, researchers, and tire kickers a way to inspect the concepts and truly experiment. License should be permissive, but one that respects and addresses the question of patents, such as Apache 2.0 License. I'd recommend a quick and dirty implementation written in one of the "easy" or "cool" languages that would satisfy all of the constituents but wouldn't be something to use to build a high performance product. See: Python, Swift, D-lang, mono, etc.