From: Leap Second Festival 
Subject: Re: proposal for leap second festival 2015
Date: June 25, 2015 at 1:53:41 PM GMT+2
To: Zsolt Mesterhazy 

Hi Zsolt,

This seems very interesting as it probes the technical issues and problems of the leap second in 
a self-referential way. I know unix systems just repeat the last second for the leap second, which 
makes it seem as if time goes backwards. There appears to be three methods for dealing with 
the leap second: repeat, freeze, smear.

I'd be very glad to include your work in the festival programme, which will be ready tomorrow, as 
today is the last day of submissions. Please send me as soon as you can the info you want me to 
include in the programme, title of work, artists, and as "shell script performance/network software 
performance" and any other relevant info. I hope you'll be able to send documentation of the 
performance afterwards to be included on the festival website. You might find out from this 
that The Leap Second Festival in reality, and not only in concept, was more of a non-event! 
(since the system didn't recognize the leap second :)

Anyway, a most interesting proposal, looking forward to hear from you,

Bjørn



On Thu, Jun 25, 2015 at 1:21 PM, Zsolt Mesterhazy ‪‬ wrote:

Hi,

How about a 'shell script performance’ or a 'networked software performance’?

We will create a script which will run exactly during the 26th leap second adjustment. The script 
will perform a http request on the leap second festival website noemata.net/leapsec26, requesting 
a non-existent document. This act will (likely) be documented in the error log file of the web 
server running noemata.net website and we will also capture the entire
interaction from one of the passing network nodes (routers) on our side.

We will reveal name of the requested document only during the performance.

NB. It's also worth elaborating how the leap second looks from the perspective of computer networks. 
The clocks in computers are being synchronized using the Network Time Protocol (NTP). The way it is 
implemented, the time in computers actually stands still during the leap second, so the time 
23.59.59 literally occurs twice and then the clock moves forward towards 00.00.00.
Also most of the computers do not synchronize their clocks multiple times per second. These will 
actually not notice the leap second at all. Most likely it will simply jump 1s forward at some 
arbitrary moment after the leap second has been inserted, discovering its clock
is suddenly 1s behind the network time.


Sources:
[1] https://www.meinbergglobal.com/english/info/leap-second.htm
[2] http://doc.ntp.org/4.1.0/leap.htm

best regards,

Radovan Misovic and Zsolt Mesterhazy