On 10/18/2021 4:04 PM, Brian E Carpenter wrote:
I think the original concern was indeed standards that (for proprietary or other reasons) were actually kept secret. So "freely" didn't imply "free of charge"; it meant available to the general public. In that sense it's closely related to "open standards". Those are standards that are open to the general public. I think that's what we insist on, and "free of charge" is desirable, but not essential. "Open standards that are openly developed" means standards whose development process is open to the general public. We don't insist on that for external references.
It's possible I missed someone else referencing this, but RFC6852 is probably a good reference for this discussion.
We have "freely available" which really means "accessible to all" not "available free of charge".
We have open standards which may (or may not) mean "processes are open to all interested and informed parties" but generally does mean "non-proprietary" and "implementable by all" and probably means "collectively created by parties with varying goals". All standards organizations - including the IETF - have barriers to entry by individuals. For us it's just that the barriers are more cultural than formal or even financial. Compare and contrast with the ITU that votes by country, or OASIS which has both organizational and personal memberships, but votes by individual. I find it hard to say where the line would be drawn between a common understand of what open and non-open standards organization processes are.
We have "voluntary adoption" which contrasts with things like NIST FIPS documents which are directive on the US government, can be directive on US government contractors, and in many cases tend to be voluntarily adopted by other groups and organizations.
Nuances are important here I think. For the purpose of this discussion, the fact that some documents may need to be paid for by the general public shouldn't disqualify them as references. That said, leaving it to the document authors to negotiate with the external standards organizations for access, or worse to pay for access, is probably not in our best interests.
Perhaps an updated version of 6852 where we can get MOAs with various organizations that we're willing to do external references to and an agreement that they will provide copies of the referenced standards for our processes might be in order. As would an agreement for stable references for archival retrieval of referenced documents. That would be a change to what Murray has currently drafted for section 6 of the document.
Later, Mike
Regards Brian Carpenter On 19-Oct-21 02:33, John C Klensin wrote:Hi. In looking through the new -01 draft (even though this text has not changed) I noticed something that I sort of hinted at yesterday in responding to other comments. You need to define "freely available" and do so precisely. We have historically considered printed books and articles in established journals to be suitable for normative references from the RFC Series ("down" really has nothing to do with that criterion) even if buying the book or obtaining the journal was expensive. In theory, there was always a trip to the library. Some of the standards from other SDOs have the same property: they are often very expensive unless one's organization is a member that gets them for free, but many libraries and other repositories do have them available. Of course, some of us have access to better technical libraries than others. That is an economic and cultural problem I don't know how to fix, but I'm fairly sure that pushing in the direction of "must be available online, with no restrictions and no cost" would be quite self-destructive for the IETF. "Freely available" does not necessarily imply "free" (zero cost). By contrast, one can imagine a reference to a restricted corporate document, some types of prepublication drafts, and, if the world continues to fragment, even the detailed description of how some equipment operates. In those cases, the document may just not be "available" to many IETF participants even though, if someone were allowed to access it, it would be at no cost. So the I-D should be very clear about what it is talking about. Then, if needed, we can have a better discussion about the requirements. best, john
