[ih] Design choices in SMTP

[Johan Helsingius]

On 07/02/2023 18:52, Dave Crocker wrote:
> At initial connection, Ed's protocol had a called site send a list of 
> its interpreted characters, so the caller wouldn't use them as data. The 
> goal was to give the caller the widest array of characters that could be 
> sent as-is, without needed to recast them in hex, or the like. (I don't 
> remember which of the usual schemes he used.)

Interesting! I wonder how much more efficient that was than the usual
schemes of using just non-control ASCII characters.

	Julf