below...
Stephen Hayes (TX/EUS) wrote:
One of the goals of re-engineering the technical publication
process is to solve the problems with providing references to
external organizations. Outside organizations often want to
reference IETF specifications, but are frustrated by their
unavailability since they cannot be formally referenced until
they exit the RFC queue.
Three approaches have been proposed to address this issue: 1.
Expedited handling - moving some critically needed documents
to the top of the queue
o Current Req-EXPEDITE-1 - The IETF technical publisher shall
expedite the processing of specific documents at the request
of the IESG.
2. Early allocation of permanent identifiers - this allows
outside documents to complete their specifications (populated
with the correct references) with the expectation that the
document will eventually pop out.
o Potential Req-PERMID-2 - The IETF technical publisher
should permit early allocation of stable identifiers for or
by the IESG to satisfy referencing requirements of external
bodies.
3. Tighten the overall publication timelines so there is not
a long wait before publication
o Potential Req-TIMEFRAMES-1 - The IETF technical publisher
should have a goal of an average publication time of no more
than 4 weeks and at least 98% of all documents should be
published within 8 weeks. Documents held up due to references
or due to a protocol action should be excluded from this
statistic.
I would propose that we focus on approachs 2 & 3 and drop 1.
I don't think we can drop 1. Even if we reach perfection with
2 and 3, I'm not prepared to assert that we will never need
expedited publishing. So it should be part of the statement of
work, even if we very rarely need it in the new world, IMHO.
Approach 3 by itself is hard to guarantee (especially with
the references loophole). If an identifier can be provided
early, with an expectation that the final document will
appear within a reasonable time, then I suspect most external
organizations will be satisfied. Most standards
organizations will accept the risk that the published
document could be delayed or possibly never appear.
There is a *reason* for requiring references at this time. We
don't want to generate citations of documents that may not be
released from the reference state for an indefinite period.
That would miss the point of why we hold documents for
normative references in the first place. Ditto if it's waiting
for an approval.
I would also propose to tighten up the requirement of the
allocation of stable identifiers from 4 weeks to 1 week and
remove the exclusions for references and protocol actions.
I agree to shorten the target time but I don't agree about
the exclusions.
o Potential Req-TIMEFRAMES-2 - The IETF technical publisher
should have a goal of an average stable identifier allocation
time of no more than 1 week and at least 98% of documents
should have a stable identifier allocated within 2 weeks of
approval.
As in my previous message, I think the actual definition
of the target times is a matter for the RFP. In this case,
the 98th percentile seems attainable, but if I was responding
to an RFP, I'd price it up to pay for vacation and sickness
cover. So even here, I'd suggest the 90th percentile.
Brian
_______________________________________________
Techspec mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/techspec
