[ntpwg] Antw: Re: Control Messages Protocol for Use with Network Time Protocol Version 4

[Ulrich Windl]

Hello!

IMHO anything that is part of a packet on the network should be documented. Also if they are neccessary for understanding the algorithm, they should also be documented. And preferrably a specific code shouldn't change it's meaning over time; it's better to use a new code instead.

I'm not deeply in NTP at the moment, but felt to comment.

Regards,
Ulrich

>>> Danny Mayer <mayer at ntp.org> schrieb am 28.12.2011 um 02:32 in Nachricht
<4EFA71B2.1070605 at ntp.org>:
> Dave,
> 
> I don't think that event codes should be documented in an RFC. They
> belong in the reference implementation documentation. They are specific
> to the reference implementation but other implementations need to be
> doing the own thing in this regard. I'm sure chrony doesn't use them,
> for example. If the status words are standard then they belong in the RFC.
> 
> Danny
> 
> On 12/26/2011 12:13 PM, David L. Mills wrote:
> > Danny,
> > 
> > This is the first I have seen the document, but my name is on it, so I
> > should respond.
> > 
> > The document  from which this is derived is twenty years old, although
> > it has been updated for recent developments. However, the system per and
> > clock status words do not align with those defined in the current ntpq
> > and status/event codes defined in the current reference implementation
> > documenttation. I suggest the description of the status words in the ID
> > be changed to agree with the documentaiton or the documentatino and
> > implementation be changed to agree with the ID.
> > 
> > The intent in the event codes is to align them with the protostats
> > statistics file messages. These are define din the  header files. While
> > these change from time to time, some kind of reference should be made to
> > the current list. I don't think this should be the IANA, so maybe the
> > reference should be to the current documentation.
> > 
> > Dave
> _______________________________________________
> ntpwg mailing list
> ntpwg at lists.ntp.org 
> http://lists.ntp.org/listinfo/ntpwg 
>