"Adam M. Costello" wrote:
> Let me make sure I'm understanding you correctly. For a given domain > name slot (protocol element, structured data field, function argument, > etc), the governing specification says The fields that are used in a message have different behaviors than the loosely-coupled data-types which appear in multiple messages or in the message payload. > what you may write into that slot, yes for both > and tells how to interpret what you read out of that slot, yes for both > but it does not dictate what you may do with the name after you've > read it. True for message fields, false for the loosely-coupled data-types which are independent of a particular of any particular protocol message (email addresses, Message-ID, URLs, etc). Also true for connection identifiers. > Do we agree on all that? No. -- Eric A. Hall http://www.ehsco.com/ Internet Core Protocols http://www.oreilly.com/catalog/coreprot/
