I don't think people should be thinking
reference C code
as opposed to
specification
All widely-successful procotols have multiple independent
implementations (sort of by definition), and that really requires a
written spec. There is another class of things which are defined by a
singleton implementation.
Certainly a reference implementation is a great thing also. That means
really clean code that does exactly the protocol, no more and no less,
really easy to follow, with really great comments. And written in a
very portable manner. And then probably with test vectors.
Right now my impression is that we don't have either, and people are
working toward both. I don't mean to sound critical -- all of this is a
ton of work, and the people doing work have accomplished a vast amount
already. I am really just reacting to "we don't need a specification".
73 de n1dam
_______________________________________________
Freetel-codec2 mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freetel-codec2
