ml>  10>> "BinaryAddress"
 AH> You've defined an integer (-32768 - 32767) for Zone, Net, Node and
 AH> Point. Why not use a word (0 - 65535) for this?

I can't believe anyone still thinks there's any difference, apart from how
local software chooses to _represent_ exactly the same 16-bit values?

JS> Because many applications use the negative numbers for "special
JS> purposes" like as a flag to indicate the presence of
"ALL" in that
JS> field.

Is 0xffff equal to '65535' or to '-1'?  Parse either, what do you get in hex?

What's the difference between -32768 and +32768, in 16-bit terms?  Zero.

 ml> can you name them and keep in mind that they should also be ones that
 ml> are actively used today

Both signed and unsigned representations are used today, with no apparent
problem (unless you're planning doing arithmetic on net/node numbers, like
echomail bundle names, where the unsigned view might give more convenient
results, though it's still the same in the lower 16 bits).

 ml> and *WILL* be updated to support your type-10
 ml> proposal?? i know that there are several packages out there in use today
 ml> that will/can never be updated as their original authors are long gone
 ml> and no source code is available for them...

What's to update?

Ian

---