RE: Header in TCP Mapping

[Robert Burbidge <robert.burbidge@poptel.coop>]

I have two conflicting responses to this, based on what hat I'm wearing at
the time.

1. Speaking as a purist, it seems sensible to change the length byte as you
suggest. When implementing the first prototype of our EPP server, my
datagram lengths were out by 4 because it simply did not occur to me that it
would be implemented as described in current drafts.

2. Speaking as a pragmatist, I would suggest leaving things as they are.
There's no real benefit in the change except from the purist point of view,
and I'm in the final stages of getting a server up right now. If it changes
in the next draft, then I and my clients will have to make one more change
to live software with no net benefit. I'm all for improving the
specification, but please bear in mind that there are several
implementations based on drafts out there already. I realise that when EPP
gets finally approved, we will all be obliged to upgrade to the final
release, but let's keep things as simple as possible in the meantime. 

Rob Burbidge
Poptel

-----Original Message-----
From: Hollenbeck, Scott [mailto:shollenbeck@verisign.com]
Sent: 12 August 2002 13:13
To: 'ietf-provreg@cafax.se'
Subject: Header in TCP Mapping


I've received a private comment about the EPP header as currently described
in the TCP mapping draft.  Here's what it currently says:

"Total Length (32 bits): The total length of the EPP datagram measured
in octets.  The octets contained in this field MUST be included in the
total length calculation."

and here's what I was planning on changing it to:

"Total Length (32 bits): The total length of the EPP data unit measured
in octets in network (big endian) byte order.  The octets contained in
this field MUST be included in the total length calculation."

The comment (really a question) was along the lines of "why do we need to
include the 4 octets for the length field in the total length calculation?"
I'm not sure that there is any real value, but before I change the wording
I'd like to ask if anyone does see any value in always adding those 4 octets
to the total length.  In short, if we have an XML instance that contains 200
octets, is it better to have the length field say "200" or "204"?

-Scott-