BCP97bis and "freely available"

[John C Klensin <[e-mail]>]

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