In recent discussion of a proposed replacement of a BCP RFC,
a couple of problems have reappeared:
1. There seems to be a fairly wide misunderstanding of what
BCP RFCs are supposed to cover. Part of the problem is that
"Best Current Practice" isn't a terribly good name for the
sort of administrative procedures and policies that BCPs
actually address. Many individuals apparently believe that
discussions of how to administer user accounts and the like
are suitable for BCP. It is clear from the RFC 2026 discussion
that that isn't what BCP RFCs are about -- for those who bother
to read 2026. Reinforcing the misinterpretation are comments
referring to "Next-Best Current Practice" and/or "Worst Current
Practice". I suspect that there would be some resistance to
changing the term "BCP" itself, so the only solution to this
problem seems to lie in better education w.r.t. the true
purpose and scope of BCP.
2. There seems to be a broad and deep lack of understanding of
and appreciation for the importance of backwards compatibility.
In searching the entire on-line collection of RFCs for an
authoritative definition and in-depth discussion of the issue,
I found none. I believe the IAB could provide a much-needed
service to the Internet Community by developing such a
definition and explanation, possibly including it in a revision
of RFC 1958, ideally with BCP (rather than Informational)
status.
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf