> Actually, the SunOS talk/talkd was also compatible with the VAX.

Well, with a VAX running the really old talk protocol.

> The idiot who designed "talk" at Berkeley sent out binary data in
> VAX-byte order (backwards from just about everything else in the
> world, even Intel),

...huh?  VAX and Intel are both little-endian; in what sense is VAX
byte-order backwards from Intel?

But yes, the basic problem is that the old talk code sent out
multi-octet quantities without converting to a well-defined byte order
first.  Since a VAX was the de-facto machine of the world at the time,
this amounted to defining the protocol to use little-endian on the
wire.  The big problem was that simply recompiling the Berkeley code on
a big-endian machine would give you big-endian data.  Sun jiggered
their code to try to do some kind of right thing in most cases, but
like most automated guesses, it guessed wrong on occasion.

Fortunately, old talk is dying off....

					der Mouse

			       mouse@rodents.montreal.qc.ca
		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B