Familypedia
Register
Line 58: Line 58:
 
Try to give an estimate, based on sibling's birth year or marriage year. If the question-marks are really a problem, even a "bef...." based on any known or estimated event in the family would do. — [[User:Robin Patterson|Robin Patterson]] [[User talk:Robin Patterson|(Talk)]] 23:42, March 7, 2010 (UTC)
 
Try to give an estimate, based on sibling's birth year or marriage year. If the question-marks are really a problem, even a "bef...." based on any known or estimated event in the family would do. — [[User:Robin Patterson|Robin Patterson]] [[User talk:Robin Patterson|(Talk)]] 23:42, March 7, 2010 (UTC)
 
::Or one could just omit the question mark, such as "John Doe (-1901)" [[User:DennisDoty|DennisDoty]] 00:03, March 8, 2010 (UTC)
 
::Or one could just omit the question mark, such as "John Doe (-1901)" [[User:DennisDoty|DennisDoty]] 00:03, March 8, 2010 (UTC)
  +
  +
:::Well, that would be a change to the standard, which should be debated. [[User:Thurstan|Thurstan]] 03:26, March 8, 2010 (UTC)
   
 
===Work-around===
 
===Work-around===

Revision as of 03:26, 8 March 2010

Forums: Index > Watercooler > Standards on dateless individuals



It seems to me that with support for the prefixes after and before, that there should not be many cases of articles like Cunigunda of Laon (?-?). Cunigunda had a child and a spouse with dates. Is there any generally accepted practice that would permit conventions such as these:

Birth

Child birth date known: The earliest known age of motherhood is 3YO, but motherhood is exceptionally rare for 10YO or younger. Guidance is therefore that if earliest known child birth year is 1500 then mother's birth year we can confidently assume is before 1490 (bef1490).
Spouse birth date known. Using the same assumptions, the spouse would have to be at least 10 to marry who we can reasonably assume to also have been no less than 10. If husbands birth year is 1450 then can't we confidently state the supposition that the wife's birth is after 1460?
We can't - she could have been much older than he. Seldom more than 10 years older, but 20, 30, 40 possible, so we could make a reasonable assumption and make her "aft1430". In my personal file she would be "c1450" until more evidence arrived; that has the advantage of listing her near her husband in a chronological list. — Robin Patterson (Talk) 00:19, December 5, 2009 (UTC)

Death.

Child birth date known: Mother's death could be at conception, so with a child birth of 1500 we know that her death had to be aft1499. Fathers could die as much as 9 months prior to child birth, so death would be aft1498.
Spouse death date known. If spouse death date of 1500 then death would have to be aft1499.

Aren't those reasonable bits of guidance on dates? I am uncertain on genealogical best practice. If these are out of line, please propose alternative language with examples. ~ Phlox 17:50, November 22, 2009 (UTC)

Dateless individuals are indeed a pain, particularly if they have the same or similar names.
The above suggestions would go awry, though, when done over multiple generations. Mother is at least 10 years, grandmother at least 20, great-grandmother at least 30 years older, but she may as well be 75 or 105 years older ...
A better alternative would be to use fl (for floruit) which just means that someone was alive around that time. rtol 20:09, November 22, 2009 (UTC)
Would that allow us to ban the question mark for dates? ~ Phlox 22:18, November 22, 2009 (UTC)
Yes. John Smith (fl1500) would have lived around 1500. John Smith (fl1298-fl1310) would certainly have been alive between 1298 and 1310 without implying that the former is close to his date of birth or the latter to his date of death. Maybe John Smith (f1500) is better, as I don't like "l" and "1" so close together. rtol 06:16, November 23, 2009 (UTC)
"l1" is an issue while editing, not while reading. rtol 06:17, November 23, 2009 (UTC)

What would be the birth year and death year for John Smith (f1298-f1310)?

I'm not sure naive users are going to grasp the distinction between circa and floruit.

Do any sites use floruit using f or fl abbreviations when refering to individuals? If so, where. ~ Phlox 06:54, November 23, 2009 (UTC)

Birth and death dates would be empty????
Wikipedia has an entry on this [1] rtol 07:02, November 23, 2009 (UTC)
I am of two minds about this. Now, if the person did not have a known birth or death year, they will not be excluded from a report, but the report lists them at the end with the value 9999. We should be able to do better than that. For the purpose of reports and other queries, we need a field that will nearly always have a value indicating the time period of the individual. Maybe floruit gets us nearer that goal because we can always take a general guess at when a person lived. OTOH it is jargon, will not be understood by the general public and is something that solves a problem for the comparatively more rare case when we know names but cannot state dates very reliably with bef, aft and circa. It's like a solution for 5% of the problems dictating what 95% of the other cases do. This is generally undesirable no matter how well motivated in principle. That is, unless you want to sacrifice appeal to lay users for the sake of technical accuracy/ purity. It can be a "religious" question, but I generally take the side that favours the general wikipedia goal of allowing everyone to be an editor. We want to make it super easy for folks to enter a few ancestors and not intimidate them with prerequisites of knowing technical concepts in genealogy. ~ Phlox 16:49, November 23, 2009 (UTC)
Fair enough. If we want to stick to c/aft/bef, then I would use circa. We already compute the age of the mother at birth, the age of the father at birth, and life expectancy at birth. So, "all" we have to do is compute the average per century (export the lot as .csv file and do this in Excel) and use this to impute the missing values. The advantage of averages (circa) is it is robust to repeated application. Minimum (after) or maximum (before) operators quickly diverge. rtol 18:16, November 23, 2009 (UTC)


I'd settle for the "c"; even a reasonable guess at "bef" or "aft" could be wrong. But for the rare cases of people about whom only one item is known - e.g. signing a document - we could use "fl" so that it's not confused with birth date. And I support the idea of cutting out the "?" because I guess it can't be handled by all systems we need to relate to, such as email. Finally, in response to the first paragraph here, I repeat that there should not ever be a page name ending with (?-?) because of the internal problems it can cause - ask Thurstan if you've forgotten. Any page like that should redirect to a hndis disambiguation page with no dates. — Robin Patterson (Talk) 00:34, December 5, 2009 (UTC)

Technical problem with question marks

The routine that creates sensor pages truncates the page name when it encounters a question mark. Thus "Edward Doty (1705-?)" becomes "Edward Doty (1705-" See Forum:Problem creating sensor page. rtol says this is a hard problem and suggests we finish the discussion here. It would surely solve the problem if we adopt a naming standard which doesn't use question marks. If we do that, someone must create a program to move all of the existing pages with question marks in their titles to pages using the new standard. There are probably thousands of pages with question marks in their titles, far too many to move manually. We must also consider whether the ancestor tree and descendants templates will operate with the redirects which will be created by moving all those pages. By the way, the program that created sensor pages for all existing person pages had no problem creating sensor pages for the pages that had question marks in their titles. There is no problem updating those existing sensor pages with question marks in their titles. DennisDoty 20:56, March 7, 2010 (UTC)

One problem at a time. What would you do with the question mark for Edward Doty?
I would say: English men in the 18th century died on average at the age of 45. Therefore: Edward Doty (1705-c1750) it is. rtol 21:58, March 7, 2010 (UTC)
Well, okay, I could do that. But I'm not happy with that solution because to me c1750 means near 1750, say 1747-1753. It implies greater certainty about the date than I feel. I am one of those naive users mentioned above who is confused with formulas for computing somewhat probable dates based on historic life expectancies. I would rather just move him to Edward Doty (1705-). It's hard enough getting users to follow any standard; if you change the standard, some of us will just stop trying. DennisDoty 23:50, March 7, 2010 (UTC)


Death year unknown

I suggest that "-?" be simply omitted. Or replaced with an estimate if something better than a "century/sex/country average" is available. — Robin Patterson (Talk) 23:42, March 7, 2010 (UTC)

Amen, Robin! I agree with simply omitting the question mark, although I would leave the dash to indicate the date is a birth date. A system based on life expectancies is just too complicated. Are we to have a table in which to look up life expectancy? DennisDoty 23:58, March 7, 2010 (UTC)

Birth year unknown

Try to give an estimate, based on sibling's birth year or marriage year. If the question-marks are really a problem, even a "bef...." based on any known or estimated event in the family would do. — Robin Patterson (Talk) 23:42, March 7, 2010 (UTC)

Or one could just omit the question mark, such as "John Doe (-1901)" DennisDoty 00:03, March 8, 2010 (UTC)
Well, that would be a change to the standard, which should be debated. Thurstan 03:26, March 8, 2010 (UTC)

Work-around

Users may be able to manipulate the address bar to replace question-marks with the "code". In a recent forum I reported on success with that. — Robin Patterson (Talk) 23:42, March 7, 2010 (UTC)