Date Descriptor

There is a need for a concise (but also precise) way of expressing approximate dates and date ranges, especially one suitable for working with medieval documents. My main application area is entering place-name spellings in a database. I am not aware of any standard defined by historians, except the s.xii² notation, and this is not powerful enough for the demands of place-name research. With the system described below, if I read that the date of a charter is “early thirteenth century to circa 1240”, I just write e13C‑c1240 and do not have to think or interpret further. No information is lost, and my outputs will be consistently formatted whatever the editor wrote. See many more examples below.

Design criteria

1. Quick to type, with no redundant keystrokes required
2. No loss of information, and no extra assumptions imposed
3. Unambiguous, but still with some looseness allowed in the inputs (so one does not have to remember too many rules; see 1W1 examples below)
4. Compact, to save space in e.g. long lists of field-names
5. Human-readable (or at least human-understandable with minimal need to refer to reference tables.)
6. Suitable simultaneously for e.g. making quick notes in the record office, for permanent records in databases, and for final use in printed and web publications.
7. Machine-readable, in a strict sense, so that invalid specifications will be detected and rejected, and also so that code can be written which “understands” the date specifications and can process them meaningfully.
8. Able to express all date information as commonly used in editions of documents (thus not requiring any re-interpretation of what an editor has already decided)
9. Automatically expandable to a more readable form (see verbose output below)
10. Regnal years handled automatically
11. Covers exact dates as well as approximate ones, and ranges with possibly both ends uncertain
12. Automatically sortable into date order
13. Uncertain dates are sorted by the latest likely date, so that nothing is listed too early
14. Case-insensitive (except for monarchal names in regnal years); though lower case is preferred for the prefixes e for ‘early’, m for ‘mid’, and upper case for C for ‘century’, and this is provided by the normalized output

Typeset example

The screenshot shows the system in action specifying some Dodnash charters for Cattawade in Suffolk, the output of the python code below having been fed through the LaTeX typesetting system.

Software

The following two files are a complete implementation in python, a free language available for all operating systems; the code works in the current python 3.5 versions, and also in the legacy python 2.7 series. The codes are offered without any guarantees: date_descriptor_09.py and regnal_year_03.py. I would appreciate bug reports or other feedback. The following extensions are in in progress:

1. A syntax t.Ed3 for “tempore” will be added soon, indentical to the standard regnal year syntax, but allowing t or t. in place of a numeral. The verbose output will expand to the limits of the reign in question, and the sort value will be at the end of the reign.
2. At present “n.d.” for ‘no date’ can only be achieved by a prenote; e.g. [n.d. ]?1435. A special syntax is being considered for this.

Grammar

An instance of DD (Date Descriptor) is parsed by first checking whether it is a regnal year by a collection of special rules. If it is not, it is parsed by the following grammar. A year value for sorting is computed (a value of -1 indicates a syntax error in the input), and both normalized and verbose outputs are also created. The syntax is best understood by looking the examples below. Note that the traditional a. for ‘ante’ is completely avoided (partly because of the danger of confusion with ‘after’) in favour of <, and the symmetrical > is provided for ‘after’. The regular expression syntax is standard (see e.g. here); for example \d means a single digit, ? means ‘optional’, [] means a single one of the enclosed characters, and $ means the end of the expression. The current version does not check for nonsensical constructions like 1234-1150, but checks could easily be added.

prenote    ='(\[(.*?)\])?'        # arbitrary text in [...]
postnote   ='(\[(.*?)\])?'        # arbitrary text in [...]
circa      ='((c\.?)|(circa))'    # "c" or "c." or "circa" for 'circa'
uncertain  ='(\?)'              
ba         ='([<>])'              # before or after
third      ='[123]t'
quarter    ='[1234]q'
eml        ='(em)|(ml)|[eml]'     # e=early, em=early to middle, l=late etc.
prefix     ='((%s)|(%s)|(%s))'%('%d',third,quarter,eml,)
century    ='(\d\d?)[Cc]'         # e.g. 12C
decade     ='(\d{2,3}0)s'         # e.g. 1260s
year       ='(\d{1,4})'
simplerange='((1\d\d\d[-]\d)|(1\d\d\d[-]\d\d)$)'     # e.g. "1243-7"
oldstyle   ='((\d{3}[012345678]/\d)|(\d{3}9/\d{2}))' # e.g. "1355/6"
first      ='{uncertain}?{ba}?{circa}?{prefix}?({simplerange}|{oldstyle}|{decade}|{year}|{century})'
second     ='{uncertain}?{ba}?{circa}?{prefix}?({oldstyle}|{decade}|{year}|{century})'
dd_grammar =prenote+first+'((([x-–])|([-]{2}))'+second+')?'+postnote+'$' # the full grammar!

The syntax illustrated by examples

Here the first column is input, and the next three columns are output from the python code. The normalized output is intended to appear in publications derived from the input; this then ensures consistency of layout and formatting. Note that an ordinary hyphen (-) for a range is converted to an en-dash (–) in the normalized output.

input               	sort	   normalized output	verbose output
--------------------------------------------------------------------------------
1234-5              	1235	              1234–5	1234–5
1234-50             	1250	             1234–50	1234–50
1101/2              	1102	              1101/2	1101/2
1109/10             	1110	             1109/10	1109/10
<1255               	1255	               <1255	before 1255
>1255               	1255	               >1255	after 1255
<c1255              	1255	             <c.1255	before circa 1255
1250-c1255          	1255	         1250–c.1255	1250 to circa 1255
1250-<1255          	1255	          1250–<1255	1250 to before 1255
1250-<c.1255        	1255	        1250–<c.1255	1250 to before circa 1255
1260s               	1260	               1260s	1260s
?em13C              	1375	              ?em13C	perhaps early to middle 13th century
em13C               	1375	               em13C	early to middle 13th century
ml15C               	1600	               ml15C	middle to late 15th century
1234                	1234	                1234	1234
cM13C               	1366	              c.m13C	circa middle of the 13th century
cm13C               	1366	              c.m13C	circa middle of the 13th century
e13C                	1333	                e13C	early 13th century
M12C                	1266	                m12C	middle of the 12th century
?M12C               	1266	               ?m12C	perhaps middle of the 12th century
1q15C               	1425	               1q15C	first quarter of the 15th century
1t15C               	1433	               1t15C	first third of the 15th century
1q15C-l16C          	1600	          1q15C–l16C	first quarter of the 15th century to the late 16th century
l15C-1q16C          	1525	          l15C–1q16C	late 15th century to the first quarter of the 16th century
9C--e10C            	 933	             9C–e10C	9th century to the early 10th century
1234-1267           	1234	           1234–1267	1234 to 1267
1234x1267           	1234	           1234x1267	1234 to 1267
c1230               	1230	              c.1230	circa 1230
12C                 	1200	                 12C	12th century
c950                	 950	               c.950	circa 950
cm9C                	 966	               c.m9C	circa middle of the 9th century
Ce16C               	1633	              c.e16C	circa early 16th century
cl12C               	1300	              c.l12C	circa late 12th century
c.l12C              	1300	              c.l12C	circa late 12th century
CL12C               	1300	              c.l12C	circa late 12th century
1 Henry 2           	1155	          1 Henry II	1 Henry II (1154/5)
3 Henry IV          	1402	          3 Henry IV	3 Henry IV (1401/2)
3H4                 	1402	          3 Henry IV	3 Henry IV (1401/2)
3Hen4               	1402	          3 Henry IV	3 Henry IV (1401/2)
3 William II        	1090	        3 William II	3 William II (1089/90)
3W2                 	1090	        3 William II	3 William II (1089/90)
1W I                	1067	         1 William I	1 William I (1067)
1 W1                	1067	         1 William I	1 William I (1067)
1W 1                	1067	         1 William I	1 William I (1067)
1W1                 	1067	         1 William I	1 William I (1067)
1W2                 	1088	        1 William II	1 William II (1087/8)
26E3                	1352	       26 Edward III	26 Edward III (1351/2)
26Ed3               	1352	       26 Edward III	26 Edward III (1351/2)
26Edw3              	1352	       26 Edward III	26 Edward III (1351/2)
26 Edward 3         	1352	       26 Edward III	26 Edward III (1351/2)
26 Edward III       	1352	       26 Edward III	26 Edward III (1351/2)
1R1                 	1190	         1 Richard I	1 Richard I (1189/90)
3R2                 	1380	        3 Richard II	3 Richard II (1379/80)
2R3                 	1485	       2 Richard III	2 Richard III (1484/5)
[perhaps]1205-l13C  	1305	   perhaps 1205–l13C	perhaps 1205 to the late 13th century
1205[, or later]    	1205	      1205, or later	1205, or later