Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Adding fault signalling to cls 49


I've run into another issue in adding fault signaling to clause 49. 

Clearly an ODDD should be encoded as in walker_1_0700.

Also, an O followed by 3 characters that are a mix of data and control
cannot be encoded. Therefore, it is clear that something like ODCD or OCDD
is an error and should be replaced by /E/s.

But what should be specified when a 0x5c or 0x9c occurs in a group of
control characters? Currently, we do not have 7-bit encodings assigned for
0x5c and 0x9c. There are 2 alternatives:

1) It will be considered an error. This is consistant with the behavior for
S (0xfb) and T (0xfd). We will not need 7-bit encodings for 0x5c and 0x9c.

2) It is allowed and the received control characters will be
encoded/decoded. This would require definition of codes 7-bit encodings for
0x5c and 0x9c.

My preference is 1 because it is consistant with the way we currently use
the code. That is what I plan to put into D2.0 unless I get inputs from the
reflector to the contrary in which case, like Katherine Harris, I'll weigh
the arguments presented.