I accept this concept by John, especially as this will move us forward.  I want to think it through a bit more, but for now, there's one obvious issue as stated in the Subject line.

With one obvious solution that will require a bit more care than we've asked so far of those who create foreign status objects.  First convention property wins (as plists and alists work), and it's important to remember this probably has little or nothing to do with programming except getting them in the right order in the list.  Unless and until we start integrating these SRFIs in development environments, except for handling raised errors the conventions list is about human interpretation of the properties of other keys in the foreign status object.

Another option is to have each convention mention in the Schemeregistry the handling of cases where a key under it is also in another, probably another specific convention, assuming there won't be very many cases that need such special handling.

- Harold