[Ntp] comments on draft-mlichvar-ntp-ntpv5-03

[Dan Drown <[e-mail]>]

After reading this draft, I'm glad to see these things are included:

1. the unknown leap seconds flag seems useful for embedded systems  
that don't have easy access to that information
2. the explicit interleaved mode with server cookies
3. client timestamps are switched to a client cookie
4. the removal of all modes except client/server

I have a few comments, and I think I'll break them up into multiple  
emails to make discussion easier.

This email will be on the "3. Data Types" section.

For the time16 definition "the maximum value is 32767/32768" This  
confused me on first read because I parsed it as either/or at first  
which made no sense. I had to think about it until I realized this was  
talking about a fraction. I'm proposing a change to: "the maximum  
value is the fraction 32767/32768 (almost 1 second)"

Should the log2 format be explained in more detail in this section?  
I've seen a lot of people confused about this data type, so I'm hoping  
a definition would help. Maybe something like "log2 - an 8 bit signed  
integer representing the rounded log2 value of seconds. For example, a  
log2 value of -2 would be 2 to the power of -2 or 0.25 seconds"

Should the change from NTPv4's time32 format be mentioned here?   
Something like:
"A 32-bit fixed-point type containing values in seconds. It has 4  
unsigned integer bits and 28 fractional bits. The maximum value is 16  
seconds and the resolution is about 3.7 nanoseconds. Note: this is  
different than the 32 bit time format in NTPv4 (RFC5905)"