RE: [emergency] Implementation Guide Outline

["Bob Wyman" <bob@wyman.us>]

Title: Message

    My personal feeling is that the current focus of the TC should be on fixing the problems with CAP V1.0 and getting as rapidly as possible to a CAP V.next rather than taking actions that would encourage anyone to actually implement CAP V1.0. An implementation guide is going to encourage the wrong thing.

    The problem here is that many of the agencies and implementers that may build CAP V1.0 implementations have very constrained budgets or attention spans. If a CAP V.next comes out shortly after they implement the V1.0 stuff, or, even after they begin to implement, it is exceptionally unlikely that they will be able to or willing to upgrade to the fixed specification in a timely fashion. It is also quite possible that if CAP V.next follows shortly after CAP V1.0, people who begin to work with V1.0 are likely to decide that the spec is too volatile to be trusted. Thus, efforts to get folk to build with CAP V1.0 will simply guarantee that it remains in use in the wild for many years or could end detracting from the long-term acceptance of CAP.

    With many protocols, having multiple versions in the wild is simply a nuisance, however, with something as critical as CAP, having multiple versions of the protocol could be very dangerous.

    I would strongly suggest that work on the implementation guide be targeted for release with CAP V.next. Thus, until the current issues with CAP are resolved, any development on the implementation guide should be focused on general principles, concepts, etc. rather than material which is likely to be CAP V1.0 dependent.

 

        bob wyman