[ntpwg] ntpwg Updated NTPv4 Protocol Specification/timestamp location

[STUART VENTERS]

Danny,
Dave,


Considering where we are, shipping the puppy is understandable.

The consequence of not having the timestamp location in the spec is:
  Interoperable time transfer accuracy is limited to 2 packet times on the slowest interconnecting link.
  Frequency transport is still at the full accuracy of the implementations and packet path.

This appears to be ok with the majority of current implementations.

Interoperability for new applications may be better served by waiting
  until the group has the time to carefully consider the matter
    in a future spec.

The current version of the spec is draft-8 which I assume is the puppy in question.

Since there doesn't appear to be much talk about draft-8 on the reflector,
   and assuming there are no squeaks about the 2 packet limit,
   does that say you have about completed a successful last call for draft-8?


Regards,

Stuart



David Mills wrote:
>
>I couldn't agree more strongly. Get this puppy out the door and save the 
>timestamp bickering for later.
>
>Danny Mayer wrote:
>
>> At this point in time (:)) I hope not. At best it should be a reference
>> in the references section. Revisiting the NTPv4 protocol draft would
>> require reviewing carefully the text and discussing it in the WG in depth.
>>
>>Stuart Venters wrote:
>>>
>>> Hopefully it can morph into an appendix to the NTP spec.
>>>