[LEAPSECS] operational time -- What's in a name?

[Rob Seaman]

On Mar 28, 2008, at 9:04 AM, Poul-Henning Kamp wrote:


> Even if we decided to fix time_t's little red wagon for

> good, and got the economic resources to do so, we would be very

> hard pressed to find the competent man-power to carry it out reliably.


I'm fascinated by your choice of this line of argument.  People are  
incompetent - let's give up.

Why then should we care about any of the trappings of modern  
civilization?

However complex the current worldwide system of systems comprising our  
civilization, it will only get more complex.  Whenever we choose to  
address some technological issue, it will be better to use system  
engineering best practices to attempt to constrain that complexity.


> If you compare with the other crap we put up with from computers,

> and the sheer mindbogglingness of the workarounds people put up

> with, I am sure that the disappearance of leap seconds would not

> even register on the publics radar.


System engineering is as much or more about characterizing the problem  
as about offering a solution.  Your personal surety is of little  
benefit to this process.  Risks that aren't characterized in advance  
are likely to register on the public radar (perhaps literally) in  
retrospect.


> My personal preference, would be that we create a new definition

> of time representation for computers, preferably in a binary format

> so the math gets faster and less buggy.


This is a nice discussion of options (absolutely no irony here).  My  
central point all these long years is that we have yet to even scratch  
the surface of capturing coherent requirements for civil timekeeping  
in the modern world.  Before we design a solution (or scrap the  
interim solution that we already have), wouldn't it make sense to  
figure out the full nature of the problem it is meant to solve?


> Provided we get 10 years notice of leapseconds, that timescale

> can contain leap seconds.  If we don't get at least 10 years notice,

> it should not suffer from them.


We would all be happy with all the notice we could get.

This is an orthogonal concept to the best scheduling cadence for clock  
updates of whatever sort.  It doesn't take much insight into human  
nature to think that a monthly cadence will get more productive  
attention than a decadal or millennial cadence.

Rob Seaman
NOAO