"Tony Hain" <[EMAIL PROTECTED]> writes:

> Yet you continue to insist that passing around incomplete topology
> information is not only valid, it is an inherent right of all application
> programmers for all time. It is fine for you to have that view. It is not
> fine when you use that view to prevent someone else from having a different
> view about how apps should deal with the topological reality of
> inconsistency. The network is not a flat, ubiquitous, consistent
> environment, so passing around points of attachment to the topology is a
> failing mode of operation. Insisting that we optimize for the fantasy
> scenario of flat and consistent ubiquity, is not helping us deliver useful
> technologies.

Its something I can't put my finger on, but the though keep reapparing in
my mind.

You seem to say you use address and scope to talk to other nodes.

What is scope ? An integer ? An bit string ? A lamdba that evaluated
somewhere ? How is scoped compared ? How is scoped stored on network ? What
is the lifetime of a scope ? How is a scope a and a adress passed to a
nohter node ? How is scope and addresses handled in your name to address
system ?

Love

Attachment: pgp00000.pgp
Description: PGP signature

Reply via email to