At 8:07 PM -0700 7/6/1999, John W. Noerenberg said:
Well, you have to identify the signature type, too. However the
changes sound minimal.
How many implementors are interested in supporting this? What bad
thing happens when an implementation meets a v5 packet it doesn't
know how to swallow? Will implementations be able to skip over them?
An implementation is flawed (in my opinion) if it does something horrid
with a future message. If a current implementation finds a V120 signature
today (which is either a data error, or evidence of time travel), it
shouldn't bus error, it should make some appropriate whimper. It's up to
the implementation what they do. Ignoring it is a fine response to my mind.
Other people may have other opinions. If they're implementors, they get to
do more than pontificate.