Dates

[CECSpecialistI]

Issue came up in review of 008 mapping:
Character position 06 indicates meanings of Date 1 and Date 2, which often repeat information in 26Z $c in a structured way that computers can interpret more reliably than 26X $c which includes all sorts of stuff.

• Does this data belong in RDA?
• Can we afford to lose it?
• If we can't afford to lose it, how will we map it?
  • Options include reification of statements, minting timespans, appending boilerplate to dates, specific properties
    This discussion should take 26X $c, 045, and 046 into consideration as well.

[GordonDunsire]

Yes, this data belongs in RDA.

Can we afford to lose it? Depends ...

How will we map it?

The values of $c in 260, 264 may not be timespans; e.g.

"1967, c1965"
"1968 [i.e. 1971]"
"<1981=>" // ???
"©2002"
"copyright 2005"

These MARC 21 Bibliographic examples indicate how difficult it will be to parse out a timespan from $c (sometimes).

So where $c is not a simple, obvious date, it is better to use 008 Date 1 and Date 2 and ignore $c.

"u" values can be simply converted to ISO 8601; e.g. "19uu" > "19"; "195u" > "195".

The value of 008/06 is also useful for parsing a value in $c:

"b": convert the value of $c to a standard BCE value such as ISO 8601.
"c": add "-" to value of Date 1 and use for 'has date of manifestation' (use "/" if ISO 8601 is used?).
"d": concatenate Date 1 and Date 2 separated by "/" (the ISO 8601 interval designator) ("--" can also be used). This is the equivalent of converting $c by replacing "-" with "/".
"e": concatenate Date 1 and Date 2 separated by "-" (the ISO 8601 date part separator), remove "#" and "u" characters from Date 2, and insert "-" in Date 2 between the second and third characters if applicable. E.g. "1983-06-15"; "1977-05", "1976-11"
"i": same as "d". Note that this is easier than attempting to parse $c "18--?-1890".
"k": same as "d".
"m": same as "d".
"n": not applicable.
"p": same as "d".
"q": same as "d".
"r": use Date 1 only for 'has date of manifestation'. The value of Date 2 can be transformed to a note: "Date of original manifestation: {Date 2}.".
"s": use Date 1 only.
"t": use Date 1 only for 'has date of manifestation'. The value of Date 2 is mapped to 'has copyright date'.
$u: same as "c".

[CECSpecialistI]

EDTF is an extension of the ISO. Registered with ISO, so approximate dates with question marks etc. could be used

[CECSpecialistI]

Questionable/uncertain dates shouldn't be recorded...they aren't accurate. How do we map such things?

Replace with broader Wikidata IRI's for decades as timespans?

[CECSpecialistI]

EDTF free, ISO $$

[CECSpecialistI]

attn: @gerontakos @pan-zhuo

[CECSpecialistI]

use EDTF

[CECSpecialistI]

Related discussion to start @CECSpecialistI --> To what extent do we reuse systematic IRI's for places/timespans/etc. in this mapping?

[CECSpecialistI]

Not for 008 at least.