[LEAPSECS] civil-solar correlation with TI

[Rob Seaman]

I wrote:


>> Historians looking backward want to relate events worldwide and  

>> arrange them into coherent timelines.


Zefram replied:


> Yes, they'll want the Olson database.


Precisely.  For a scheme such as this to have any chance of working, a  
requirement is that it be tightly coupled to a mechanism like  
zoneinfo.  This is equivalent to Steve Allen's proposal.


>> Whatever the preferences of the ITU, they will discover that it is  

>> simply unacceptable to allow local dates to vary secularly from  

>> civil timekeeping dates.

>

> I don't see how this follows.  Given the Olson database they'll be  

> able to apply the offsets correctly.


A further requirement is that there needs to be faith in that database  
and in how it is tied into the fabric (system of systems) of the world.


> If the date drift per se really is a problem, that would be a reason  

> to argue for the IDL-jumping version of my scenario, rather than the  

> unbounded-timezone-offset version.


Words like "jumping" and "unbounded" reflect that the discontinuities  
represented by leap seconds remain inherent in the system.  One way or  
another, intercalary corrections (of whatever sort) will remain  
necessary.  Since they are necessary, so is a coherent and reliable  
mechanism for managing them.  The devil-may-care ITU proposal is  
insufficient.

Rob