Should DB errors 555 and 558 be adapted for julian/gregorian calendar?

Question

This is inspired by the result of an error report run on King Charles VIII of France (Valois-273)

Many pre-1600 profiles with a Wikidata counterpart trigger errors 555 and/or 558, obviously because one profile uses the Gregorian calendar and the other uses the Julian one, resulting in a (roughly) 10 days difference between the dates.

I understand Wikitree policy is to use Gregorian dates whenever possible. I don't know about Wikidata. Some dates entered into Wikitree may be Julian where they should be Gregorian, but I'm not sure. Right now I tend to tag those errors False when they concern a profile before 1600 (roughly) and the difference is about 10 days, with a comment  like "Julian/Gregorian calendar ambiguity".

Could the error report be adapted to avoid raising false errors, or a more comprehensive explanation added to the help page for these two errors? I'm not sure what the appropriate solution is and of course the issue is complex since countries switched to Gregorian at different periods (beware of Wikidata dates imported from Russian Wikipedia !). Here is a list: https://en.wikipedia.org/wiki/List_of_adoption_dates_of_the_Gregorian_calendar_per_country

Thank you!

(Adding euroaristo and notables tags because of the Wikidata issue).

Answer 1

I agree 100% a genealogy website that is world wide should better support dates and different calendars

We need

• precision = how exact a date is e.g. day, month, year decade
• define a range
  • before, after....
• calendar used

I suggested in aug 2016 a quick and dirty solution that we should use a template to tell what calender is used ==> 

1. The Project database error report will understand it and can compare with Wikidata/Wikipedia  ==> No errors
2. A reader understand what calendar this date used ==> No guessing for the WikiTree reader 

See a draft of the template I created that has never been used Template:Calendar. 

Comments

In Wikidata Charles VIII de France is wd:Q134452 ==>

1. Query
2. Result
   1. That shows
      1. Timeprecision
         1. A date can be defined to be exact, month, year, decade....
      2. Timezone used
      3. Timecalendar e.g. Proleptic Julian calendar

By changing the Q number you get the result for another person e.g. Medici-19 is Q48544 ==>

1. query
2. result

A more technical approach is check the data structure how its stored in WIkidata and can be accessed....

1. Charles VII de Frances Q134452
   1. P569 is birth
   2. P570 is death

       "P570": [
                          {
                              "mainsnak": {                            "snaktype": "value",                           "property": "P570",
      "datavalue": {
      "value": {
      "time": "+1464-08-01T00:00:00Z",
      "timezone": 0,
      "before": 0,
      "after": 0,
      "precision": 11,
      "calendarmodel": "http://www.wikidata.org/entity/Q1985786"                                },
      "type": "time"},
      "datatype": "time"
                              },

2. Medici Q48544 

See also earlier discussion G2G

---

Well. Wikidata has 9 July for Charles VIII and Wikitree has 30 June. The difference is the same for his DOD.

But, if Wikitree used the Gregorian calendara and Wikidata the Julian one, it should be the other way around (i.e. the Julian calendar is roughly 10 days behind the Gregorian one - for this period). Correct?

I'm confused. One thing is certain; the pre-requisite for this to work is to know which calendar your source is using.

---

>> One thing is certain; the pre-requisite for this to work is to know which calendar your source is using.

And Wikitree should make it easy for the reader and show what calender we use.... today its some unspoken rules... that at least Project Database error and I has problem to understand ;-)

Good presentation I found a link to in G2G

---

Maybe to work out Charles VIII dates, we need to check sources closer to the time he lived?

Anselme, 3rd edition has 30 June 1470, but we don't know whether that date has been altered to match a different calendar?

---

I don't think Père Anselme says which calendar he is using, but it's very unlikely he would have converted pre-1582 dates to a proleptic Gregorian calendar though.

There is some evidence: we know for certain that his dates match those recorded by Louise de Savoie for the births of her grandchildren. Louise was recording contemporary events and could not have used another calendar than Julian. (Louise's diary, quoted in Medieval Lands, vs. Anselme's works accessed through Gallica).

I'll come back to this issue after the Clean-a-thon (my team turns out highly competitive, can't let them down !)

---

Wikidata can be a good candidate to start finding sources and get an understanding of calendar used. I have no skills in this type of genealogy.

see 

• the squid interface of Q134452 Charles VIII of France
• Wikipedia Charles_VIII_of_France#References

---

I believe you may assume that dates up to and including 4 october 1582 follow the Julian calendar. There is not much sense in converting those to proleptic Gregorian IMO.

Actually it should be up to 14 october 1582, since the first day of the Julian calendar was 15 october 1582 (Teresa of Avila died in the night from 4 october to 15 october 1582).

---

@Isabelle Rassinot 

?!?!? I dont follow you...

>>There is not much sense in converting those to proleptic Gregorian IMO

Why? If we use a calendar we must explain for the reader what calendar we are using or?

As this World Wide Family Tree you will not just have Proleptic_Gregorian_calendar or Julian_calendar you have many more see Template:Calendars

Also when WIkiTree now start comparing dates with other sources like Wikidata we must explain this is a date using calendar xxxx.....

Ps. I have seen at Wikipedia that just to make it clear for the reader they in the text use a Template that you add after the date and then the reader can click on the text and learn what date is used etc.. maybe something for Wikitree... I dont know if someone is pushing this issue....

---

I mean that all dates (I should say: in "occidental" countries, countries which used a Julian calendar at the time - I'm not talking about Asia...) up to 4/14 October 1582 were recorded at the time using the Julian system (the only one used in Occident at the time AFAIK). And the convention has remained, since then, to keep the julian as official dates. Go to the Wikipedia entries for any notable or royal of that era, you will see that the dates are Julian.

I believe translating those dates (before 15 October 1582) to proleptic gregorian on Wikitree profile pages - on the data fields - would be extremely confusing because anyone looking at the profile who is familiar with the person profiled would say "Ah! But this is wrong! He was born/died/married 10 days before that!".

I have no problem with adding a proleptic Gregorian conversion in the biography, or anything that you suggest above.

All that I'm saying is that, on a pre-1583 profile without more information, the overwhelming odds are that the dates use the JULIAN convention.

Just to make it clear, I would fully support having several calendars available, and indeed it is the only solution to make things clear for the long period (1582-1923) where Julian and Gregorian calendars co-existed.

Clear as mud, right?

---

I think we agree and see the problem. Solution is add a field next to dates where you can add the calendar used....

Having a genealogy site where people needs to guess what calendar is used is not good genealogy...

---

Dates can be stored neutrally, but if you want to use them, you want the computer to serve them up in a friendly format.

WikiData users are mostly going to want them in the real-world calendar appropriate for the place and time.

Trouble is, the calendar choice is itself a function of the date, and other things.

So the programmer is in a bind, because he doesn't know what format he wants to request the date in without knowing what the date is.

---

It's like names.  You can store somebody's name in many languages, but if you want the facility to retrieve somebody's name in their own language, you have to store which language that is.

Answer 2

In France you also have the Republic Calendar https://en.wikipedia.org/wiki/French_Republican_Calendar

Comments

Oh yes. One thing with the Republican/Revolutionary Calendar is that dates in this format won't be accepted by Wikitree. And there's no ambiguity. On the other hand, several profiles from gedcom downloads have missing dates because the revolutionary format dates were not accepted.

When entering a date from that era (in France, 1793-1806, but the calendar was also used in other parts of Napoléon's empire, like current Belgium and Netherlands, Luxemburg...) I always write out both forms (revolutionary and gregorian) in the biography.

Here is the converter I use (there are lots of others).

---

See also: https://en.wikipedia.org/wiki/Chinese_calendar

Answer 3

Another complex profile

• Распутин-1 that is same as
  • Wikipedia Grigori_Rasputin same as
    • Wikidata Q43989 same as
      • Findagrave #7304586

1. WikiTree Распутин-1
   1. Born 21 Jan 1869
   2. Died 30 Dec 1916
2. Wikipedia Grigori_Rasputin
   1. Born
      1. Old style 9 January 1869
      2. 21 January 1869
   2. Died
      1. Old style 16 dec 1916
      2. 30 december 1916
3. FindAgrave #7304586
   1. Born Jan. 10, 1872
   2. Died  c. Dec. 30, 1916
4. Wikidata Q43989
   1. Born 9 januari 1869^Juliansk
   2. Died 17 december 1916^Juliansk  

NOTA BENE:  Easy solution in Wikipedia to explain the different calenders used. What is needed

1. Create a space page to explain Old style
2. Add both dates in the Bio section with the link to the Old style page
3. Negative
   1. Can be a problem for Project Database Error to understand but that could be addressed with a template