[tz] [PROPOSED] Mention RFC 8633 on leap smearing

Hal Murray hmurray at megapathdsl.net
Wed Sep 11 21:51:57 UTC 2019

tz at iana.org said:
> I see that RFC 8633 cautions against public-facing servers' using leap
> smearing, and NTP clients' using a mixture of smeared and unsmeared.  I
> suppose the same should apply to quasi-NTP UT1 servers such as: 

Correct, but the nature of the errors is different from smearing servers.

There are many smearing servers.  There is no way in the protocol to tell if 
you are talking to a server that will smear when a leap second comes along.  
The whole point of them is to trick innocent clients into avoiding leaps.

If you use a UT1 server, the error will likely be visible right away.  If you 
happen to catch it near the zero crossing, it will soon drift off.  There 
aren't many UT1 servers.  You are very unlikely to use one by accident.

I don't think it's worth cluttering up TZ info with a warning about UT1 
servers.  It might be useful as an example of a more general warning.

These are my opinions.  I hate spam.

More information about the tz mailing list