> If this new and enhanced "standard" would be accepted, then the
> "hey-lets-mask-the-7bit" crowd would have some difficulties using it.
> (Serves them right, for making us 8th bit crowd having difficulties.)
>
> For example, RFC822, (a bit dated), states explicitly that the 8th bit
> *must* be zeroed, even if the transport mechanism below can transfer 8 bit
> octets undamaged. Why is that?
>
Well, this goes back to the time of Egil Skalagrimsson, when communication
devices used the 8th bit for something called parity. Many of these devices
were thick-skulled and sturdy enough to survive to this day. Thus, following
the well-known network principle "be liberal in what you send and conservative
in what you receive" (or was that the other way round?), some of the more
successful networking protocols have to assume there might still be an axe
swinging somewhere between the two end systems that they don't know about.
- Frank
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:37 EDT