Problem: We have a lot of specifications that work fine, but have not seen
Examples: Most security specifications, IPv6, etc. etc.
Cause: The economic case for deployment is not made. This is a particular
concern when the protocol deplends on the network effect. Until the network is
established the protocol has minimal value.
Proposed Solution: Deployment Cases
As engineers we are now used to considering use cases, that is representative
use scenarios that illustrate particular intendeded uses. The output of use
cases are requirements.
Deployment cases are similar except that we consider the economic incentives
For example, Alice the home user has three computers and two network
addressable printers, all of which currently support IPv4, Bob is the netop at
a medium sized corporation, Carol is CIO of a F500 corporation.
For each adoption case consider the costs and benefits of proposed transition
I develop the underlying theory in The dotCrime Manifesto. The problem we face
is very similar to the problem faced in the area of energy conservation, how
does a state persuade consumers to use less electricity so it does not need a
new power plant? Until this discussion however I had not really considered the
problem of how to build these considerations into the standards process.
We have to take adoption really seriously. Take a look at your email client,
chances are 0.95 or better that it supports S/MIME. How many emails do you see
each day that are S/MIME signed? We cannot accept a 'success' of that type here.
Ietf mailing list