[ppml] Policy Proposal 2007-15: Authentication ofLegacyResources

[Paul Vixie]

> >> So, you have 100 people and water for 50.
> >>
> >> It isn't a question of storage, it is of allocation.  What would you
> >> propose?  FCFS?  "To each according to need"?  A market?
> >
> > can you define more of the constraints?
> 
> Why?  The statement was that "water should never be rationed."

your question was "what would you propose?"

> Your response is asking for what would constrain the rationing, not whether
> the rationing was necessary.

my response is to learn more about your question so i can consider it.

> > i can't think of an example where rationing was used for a fixed resource
> > which was going to run completely and forever out at a predictable moment.
> 
> ?
> 
> Any limited resource, land, gold, food, water, oil, etc. are all  subject
> to rationing when the demand outstrips supply.

we're not talking about limited resources, we're talking about dead-end
resources.  last of the mohicans resources.  more fresh water will come
out of a spring or melt out of snow (for now, anyway), and so rationing
works because there's both a present allocation and a future supply to
be considered.  ipv4 has no future supply, only future zero-sum.

> The fact that a resource will "run completely and forever out" at a
> predictable (or un-) moment is irrelevant.  The point is to distribute the
> resource as "equitably" (for some value of that variable) as possible during
> the period of scarcity.

humour me.  it actually is very relevant whether a future supply will exist.
or whether it's a zero-sum game.  this changes what you call it, how you
treat it, what you can expect from it -- it changes everything.  if you'd
like an answer to your hypothetical question, i'll need more constraints.

note that at the macro level, the "soft landing" proposal had one similar
element to my open-mic ("bad ideas night") proposal which i now see echoed in
a current proposal -- get folks to describe their ipv6 plans before they can
get more ipv4 space.  noone should be deploying non-dualstack in this day+age.
(it should go without saying that i make that suggestion as a member of the
community and not as a member of arin's board -- i don't know what the board
thinks of this idea but it is unlikely to be identical to my personal views.)