Dates: complex encoding

date calendar AD and BC dates year month Gregorian calendar phrase-level encoding
date docDate dateRange

Encoding dates using date and the value attribute, including detailed instructions on the ISO8601 standard for date values

A slightly more complex encoding for dates may be needed if you are dealing with documents that use different calendar systems, or if you need to be able to express uncertainty about dates. The TEI provides a calendar attribute for date, and suggests values including Julian, Islamic, and Revolutionary (i.e. French Revolution). These could be supplemented with additional values as necessary. Calendars are discussed in more detail in Dates: dealing with calendars.

A certainty (P4) or cert (P5) attribute is also provided for cases where the accuracy of the date is uncertain. Be sure not to confuse uncertainty with imprecision; the former indicates that you suspect the date of being incorrect, where the latter indicates that precise information is missing from the date.

In the P4 Guidelines, the TEI also provides a much more complex encoding scheme for highly structured date information, described in Chapter 20.4 of P4, but this material goes well beyond the needs of typical encoding projects. In fact, it is used so seldomly it was removed from the P5 versions of the Guidelines. It allows each component of a date or time expression (month, day, year, minute, etc.) to be captured and regularized in a separate element. Such an encoding would support more complex analysis, but would only be warranted in materials for which date information is widespread and extremely significant.

Examples

Example 1 (P4): a precise date of doubtful certainty

<date value="2006-09-07T14:00-04" certainty="approx">Thu 07 Sep 06</date>

Example 1 (P5): a precise date of doubtful certainty

<date when="2006-09-07T14:00:00-04:00" cert="medium">Thu 07 Sep 06</date>

Example 2 (P4): an imprecise date

<date value="2007-03" certainty="good">early March, 2007</date>

Note that the normalization is only precise to the month, whereas the content is slightly more precise than that.

Example 2 (P5): an imprecise date

<date value="2007-03" cert="high">early March, 2007</date>

Note that the normalization is only precise to the month, whereas the content is slightly more precise than that.