Re: hrm

[Erik Ostrom <eostrom@research.att.com>]

> it just ocurred to me we probably need to make similar statements
> about cord types to the ones we make about protocol names, sigh.  any
> comments or ideas?  should the section on cord type arguments say
> pretty much the same thing the protocol section does?

I'd go a step further and link cord types explicitly to protocol
names.  I.e., just as you can only create messages that have a prefix
which you own (say, "dns-com-att-research-twin-edit"), you can only
create cord names that have a prefix which you own.  note that this
can be a non-strict prefix, e.g., if there's a "twin" protocol, we
could have a "twin" cord type and a "twin" message.

--Erik