add clarifications to design decisions on Template Distribution Protocol

[plebhash]

while the core proposal of #167 was rejected, a few reasonable points were raised with regards to the need for clarification around some design goals of the Sv2 Template Distribution Protocol

in summary, the main topics that should be more clearly articulated are:

• merged mining
• the ability to run TP as a server that caters to multiple clients

[plebhash]

I'm particularly curious about this sentence:

sv2-spec/07-Template-Distribution-Protocol.md

Line 5 in 2c21bcf

While not recommended, the template update protocol can be a remote server, and is thus authenticated and signed in the same way as all other protocols (using the same SetupConnection handshake).

As @Sjors pointed out on #167, it seems to contradict some of the design goals mentioned on that discussion.

@TheBlueMatt do you know what was the original rationale behind it? Or more specifically, why is this framed as "not recommended"?