Nathan:

date of birth should require the entire date. If just the year is requested, then it would be called "year of birth". 

 

That’s not really true. BIBFRAME for example has many date properties, all with “date” in the name of the property (none with “year”). If a date is supplied then xs:date is appended as the datatype, if a year, then xs:year.

 

e.g. the RDF triple:

<instance>   bf:dateOfPublication   “1986” ^^xs:year

 

Ray

 

 

 

From: Discussion of the Developing Date/Time Standards [mailto:[log in to unmask]] On Behalf Of Nathan Harrenstein
Sent: Monday, March 28, 2016 4:25 PM
To: [log in to unmask]
Subject: Re: [DATETIME] distinction between a period of time and an event

 

I agree with John. For date of birth, I would use the "one of a set" feature ( e.g. [ 2001-06-06..2005-10-10] instead of  2001-06-06/2005-10-10). To me an interval always means a period of time and never an event.

 

As an aside, date of birth should require the entire date. If just the year is requested, then it would be called "year of birth". 

 

Nathan

 

On Mon, Mar 28, 2016 at 1:19 PM, John Hostage <[log in to unmask]> wrote:

To me these are two separate concepts, and I didn’t realize “we” were lumping them together.  I don’t think ISO 8601 lumps them together, nor did the draft of EDTF.  The solidus (/) is used to denote a time interval.  A note to the definition in the draft of 8601-1 says “A time interval comprises all instants between the two limiting instants and, unless otherwise stated, the limiting instants themselves.”  A dateOfBirth that was recorded as 2001/2015 would indicate an extremely long period of labor for the poor mother.  For an event like a dateOfBirth we are interested in a “time point” (ISO 8601-1 sec. 2.1.5), not a “time interval.”  If you want to say it was sometime between 2001 and 2015, you can say [2001..2015].  Otherwise, why even have sec. 4.9 in 8601-2?

 

Which bibliographic applications use the same form for both purposes?

 

So I agree with the commenter, it is not the same thing.

 

John

 

 

 

From: Discussion of the Developing Date/Time Standards [mailto:[log in to unmask]] On Behalf Of Denenberg, Ray
Sent: Monday, March 28, 2016 14:57
To: [log in to unmask]
Subject: distinction between a period of time and an event

 

To some extent we have waved our hands over the semantic distinction between a period of time and an event, lumping the two together as an interval.

 

Consider two data elements, dateOfBirth and lifespan.

 

(1)    dateOfbirth

·         2001/2015 means that the person was born during one of the years 2001, 2002, …, 2015.   (Or you could say, sometime during the period beginning in 2001 and ending in 2015. Some like the first characterization because this is expressed with year precision but to me these say exactly the same thing.)

·         2001-06-06/2005-10-10 means that the person was born some day between 2001-06-06 and 2001-10-10 (inclusive).  

(2)    lifespan

·         2001/2015 means the person  lived from 2001 to 2015.    

·         2001-06-06/2015-10-10  means the person lived from 2001-06-06 to 2015-10-10.

(For lifespan, nevermind that some other application might define two elements, dateOfBirth and dateOfDeath; in this example a single element represents the span of time that two separate elements in a different application might represent the start and end of.)

There is an element of precision in both (1) and (2) but I don’t want to talk about precision now. What I want to talk about is the fundamental difference  between (1) and (2):   The first is an event; the second a period of time.     (The first, an event that takes place in the period of time represented by the interval, but an event nevertheless and not a period of time.)

 

I have always maintained that whether 2001/2015 represents a timespan or a period during which an event occurred is application dependent; it depends on the semantics of the data element for which it is a value.   8601 (Part 1) is completely silent on this issue (definition -time interval: part of the time axis limited by two instants).   Bibliographic applications use the interval form for both purposes.

 

Why do I bring this up?

 

One of the comments on the recent draft suggested the use of:

 

  “..” for representation of an unspecified value that is “between” two dates (inclusive), such as “2015-01-22..2015-04.20”.

 

And the comment said:

This is not the same thing as a closed interval of “2015-01-22/2015-04-20”.

 

I am arguing that it is the same.  (Or rather, that the interval form may be used for either, and so the suggested additional form is not necessary.

 

Comments, please.

 

Ray