"Eric A. Hall" <ehall(_at_)ehsco(_dot_)com>>
Advocating the use of a ~pointer towards the policy statement instead of
the statement itself is precisely within the charter, as the pointer
would
certainly be "associated with" that data.
So why not have the MARID record being an indirection to the data, got via
some other protocol? This might provide all kinds of useful features:
dynamic stuff, cacheable stuff, multiple content-types, 3rd party service
of the data... On the face of it - it seems really neat. I guess there's an
issue with the workload placed on a busy recipient system, clearly, it
probably isn't as lightweight as a pure DNS scheme, but local caching might
go a long way to limit the overhead. Is there more exposure to DOS on
publishers (or publisher's agents)?
OK, so it's much heavier, but some of these MARID proposals are getting far
from lightweight themselves. Presumably, this has all been considered and
dismissed already?