Apologies for not responding sooner.  I am now trying to catch up with a number of issues raised over the past  month or so. We have finally caught the attention of ISO and there will be a meeting soon to begin discussion of this spec.


On the issue (“Argument for Change in Uncertainty/Approximation”)  I think this argument has merit -- either reverse the direction of inheritance of remove inheritance altogether.   I will certainly bring this up during the ISO process.




From: Discussion of the Developing Date/Time Standards [mailto:[log in to unmask]] On Behalf Of Nathan Harrenstein
Sent: Tuesday, March 31, 2015 3:04 PM
To: [log in to unmask]
Subject: Re: [DATETIME] Argument for Change in Uncertainty/Approximation


Yes, exactly, we are saying the same thing. Approximation or uncertainty is more likely to begin at the day level (though each component is independent) and thus it should inherit from right to left, but currently it inherits from left to right.





On Mon, Mar 30, 2015 at 10:30 PM, Richard Tallent <[log in to unmask]> wrote:

On Mar 30, 2015, at 2:31 PM, Nathan Harrenstein <[log in to unmask]> wrote:

Uncertainty of a date is usually applied first at the most specific part of the date, and then to the increasingly less specific parts. (ie. from the day to the month to the year):


Uncertainty about each part is independent — for example, being uncertain about the month says nothing about the certainty you may have about the year or day.


Approximation is more likely to “inherit” from right to left, but that’s not always the case either. For example, if a record says someone was born on July 4 and is currently around 40 years old, a date of 1975~-07-04 would be appropriate.



Another possibility would be to remove parentheses, and have flags always apply to the part of the date to which it is applied.


I also find the parenthesis problematic for parsing, so alternatives like this should be brought up during the ISO process.