What is a "communication" (was Re: sorry, but ...

[Caspar Bowden (travelling)]

On 07/31/2012 11:15 AM, Roland Perry wrote:
> That's over-analysing the situation. And in any event a transmission by
> TCP/IP involves a handshake, so the message is provably partly received
> even before the transmission has complete.
>> The RIPA word is "communication" not message (my fault), and I do not
>> think the communication (as opposed to the packets or bits) is received
>> until it has all reached some destination.
> A communication is a flow of information, not a lump of information...

If I may, I wanted to fork the thread here to focus on what we mean by a 
communication, in the sense of the logical level of the protocol stack.

It always seemed to me that "communication" had to be interpreted as a 
transmission of information at an arbitrary *logical* layer of the 
stack(s) - it might mean an e-mail, or a web page, or an IM, or a phone 
call, or an SMS etc. If a hacker was communicating by port-knocking, it 
might mean a datagram.

So the interpretation of communication w.r.t. to whether something is 
internal or external, would not be affected by e.g. whether any dropped 
packets as part of an email message were received by a random router 
outside the UK (and perhaps "made available" to a engineer looking at a 
log file), but whether the (intended) sender and receiver are both in 
the UK.

That was another major motivation for the probing of s.16(3) resulting 
in Bassam's letter. The logic is that ECHR would prohibit discrimination 
by nationality (unlike US law like FISA), so the discrimination between 
mass and targeted surveillance occurs according to the internal/external 
criterion. In a packet switched network, that criterion only makes sense 
if you focus on the location of the sender and intended receiver, at any 
given logical layer. So the S.16(3) probing was intended to force the 
govt. to acknowledge "if that's what you want, then this is what is 
going to happen"

Caspar