Reconsidering whether the spec should support internal uncertainly/approximation, recall that Markus Flatscher provided use cases for the archives and digital scholarly/documentary editing community. (See http://listserv.loc.gov/cgi-bin/wa?A2=ind1104&L=datetime&T=0&X=57E82753D02E3C83FD&P=1123.)
- "7 April 2011?". The classic example - it is unclear whether it is the year only which is uncertain (and the day and month are certain) or the entire date, without consulting the editorial apparatus.
- "21? December 1783". Day uncertain, month and year certain. (The Papers of Thomas Jefferson Digital Edition, eds. Oberg and Looney, Main Series, vol. 6, p. 410)
- "11 Sep.? 1786". Day certain, month uncertain, year certain. (ibid., vol. 10, p. 350)
- "ca. 1 Mch.? 1785". Day approximate (but certain), month uncertain, year certain. (ibid., vol. 8, p. 9) [Im not sure what "approximate (but certain)" means.]
- "3-6? March 1775". An interval of days where March 1775 is known but the exact days are not. (The Papers of James Madison Digital Edition, ed. Stagg, Congressional Series, vol. 1, p. 138)
- "April-June? 1786". An interval of months where 1776 is known but the
exact months are not. (ibid., vol. 9, p. 4)
Please may I have opinions on whether these use cases justify the additional complexity, particularly from those of you who have commented that the feature adds too much complexity.