> In the future there may be > other experiments that we come up with. It's not possible to figure > out what sorts of requirements we might need in the future to > accomodate some not yet thought of experiment. If the desire is just > to allow experiments, I think it is more appropriate to add a new > requirement to section: 3.20 (Process and Document Evolution). I > think this is already covered, but if we want to make it explict, we > should add a new requirement along the lines of: > > Req-PROCESSCHG-2 - The IETF technical publisher should participate in > and support process experiments involving the technical publication > process.
I think this is much better than the initial requirement, as it no longer focuses on specific needs of a single experiment. But I think the issue we are discussing goes beyond experiments. The question is IETF's ability to affect what the publisher does. There are several aspects in this. One is the ability to run experiments, even those that go against currently approved requirements. Another is the ability of the IETF to change the requirements. Your text above captures the former. The latter is captured elsewhere in the document: > The IETF > technical publisher must be part of the discussions of these policies > and be prepared to implement and facilitate policy changes as they > are determined by IETF consensus. ... > Req-PROCESSCHG-1 - The IETF technical publisher should participate > in the discussions of changes to author guidelines and publication > process changes. Both are essential, I think. (Of course, they may result in additional negotiation over, for instance, new tasks required to be taken on by the publisher.) These requirements allow us to cover, for instance, changes to handling normative references as in draft-klensin, "service level differentiation" as proposed in draft-eronen, adoption of new input formats to the publisher, new technology that allows automation of some steps in a way that the publisher no longer needs to perform a particula task, etc. Is this all that is needed? In draft-klensin/eronen, for instance, the -2 requirement would allow running the experiments. If the experiments are successful, the -1 requirement would allow a permanent change. --Jari _______________________________________________ Techspec mailing list [email protected] https://www1.ietf.org/mailman/listinfo/techspec
