On Friday 07 December 2007 09:32, Dave Crocker wrote:
Patrick Peterson wrote:
I recommend we all agree that we don't know when these things will
happen. It certainly won't be in less than 12 months and it had better
be before "pretty much forever". I think we can still agree on the best
design even if it takes somewhere in this range to realize. We can
certainly agree that the clock won't start ticking until we have a spec.
and
Patrick Peterson wrote:
> Similar to my recommendations on avoiding crystal ball arguments in
> designing SSP, I would like to encourage us to avoid arguing about SSP
> market demand. Again this thread from Dave and Arvel is a great
> illustration.
Patrick,
Your suggestions would be fine if the standards effort were free, rather
than representing massive direct and opportunity costs, as well as
affecting the complexity -- ie, the basic viability -- of the
specification.
A sense of the market need/demand for a protocol is fundamental to the
decision to pursue developing it. A sense of the plausible adoption rate
has a massive effect on the types of engineering decisions that are made
about it.
(By the way, the issue of market demand is formally written into IETF
documents about chartering an effort. The comment about adoption is simple
experience: SSP seeks to change existing email handling. Changes to
existing service take a very long time. We have lots and lots of
experience to back this statement up.)
The difference between narrow, immediate need versus broader, long-term
need (as wells as between combinatorials of these) will typically have a
very large impact on the specific engineering choices that are made and the
way they are pursued, such as shoving something out the door without
testing it, versus taking a more gradual approach that permits learning
from taking smaller steps.
For myself, I am clear there is an immediate need among a small set of very
important service operators. Given that we all know of at least one
real-world example of a private agreement for an equivalent service -- as
well as statements from some others that they want the same service -- I
think we have a solid basis for believing that there will be uptake among
this market niche.
What we also have is, at best, conflicting data about the broader market.
It makes no sense for us to decide to ignore this fact.
d/
We had this discussion when the working group was chartered. The decision to
include SSP in the charter is a reflection of the consensus. Given that, is
there clear and convincing evidence that would merit overturning the
established consensus?
Scott K
P.S. Please could we do something new and stop redoing the last two years of
discussion on this topic.
_______________________________________________
NOTE WELL: This list operates according to
http://mipassoc.org/dkim/ietf-list-rules.html