Talk:Q1405

Dates of birth and death

I have reverted edits by User:Candalua. In the case of date of death he has caused to be stored into the database "ISO timestamp +00000000014-08-19T00:00:00Z" which is in the ISO 8601 format and therefore is a Gregorian date. The correct Gregorian date of death is 17 August AD 14, so the edit is incorrect.

In the case of the date of birth, the date 23 September 63 BC from the BBC is presumably a Roman calendar date, although the article does not say. Dates from that era cannot usually be translated into the proleptic Julian calendar or prolepic Gregorian calendar with a precision of better than a few days. I will see if I can find a date with better provenance. Jc3s5h (talk) 14:37, 10 February 2015 (UTC)

I have an online copy at the University of Chicago website of a translation of the early first century historian Suetonias, translated in the early 20th century. It confirms my suspicion that 23 September BC is stated in the calendar of the Roman Republic. It was much different that the Julian calendar; when Julius Caesar reformed the calendar he did so by decreeing that 46 BC would be 445 days long (an extension of about 2 months and 20 days). I have not been able to find a resource that converts Roman calendar dates to the Julian or Gregorian calendar (any such attempt would be an approximation). Therefore I have changed the date of birth to 63 BC. Keep in mind that Wikidata only attempts to support two calendars, Gregorian and Julian, so there is no way to represent a Roman calendar date. (And of course the Julian calendar support is broken).

My knowledge so far, after discussing with you and others, is that dates should be inserted as-they-are, provided that the right calendar is selected. The way they are stored in the database is just a convention, because only the Gregorian format is currently available. Once you label a date as Julian, that means the date has to be interpreted as Julian and not Gregorian, regardless of how it's stored in the database. Maybe in the future Wikidata should provide full support for all existing calendars, including database support, so that the date can be correctly stored how it really is. But in the end what really matters is what the users see. This date must be displayed as "19 August 14 (Julian)", therefore we should write the value so that what is displayed is 19 August, Julian. Otherwise we're just giving a false information to users. Candalua (talk) 15:30, 10 February 2015 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ I will copy some notes I made on the topic:

The original proposal for P569is at

https://www.wikidata.org/wiki/Wikidata:Property_proposal/Archive/8#P569

The proposal says the birt date it is a TimeValue.

MediaWikiWiki:Wikibase/DataModel says TimeValue has a time thus:

time (isotime): point in time, represented per ISO8601, they year always having 11 digits, the date always be signed, in the format +00000002013-01-01T00:00:00Z

Pasted from <https://www.mediawiki.org/wiki/Wikibase/DataModel>

It also has a precision, after, before, and timezone parts. Then it has a calendarmodel, which should be used to display the time value.

Finally there is a calendartime, which is the time in the format specific to the calendarmodel field.

Another way of looking at it is that the user interface you get when you enter something in the search box is only one way of viewing data from Wikidata. There are several other mechanisms for extracting data. All of them must know what the convention is for the date they extract, and the convention is as I just described. The fact that the user interface presents the data in a misleading way is not an excuse to store convention-defying information into the database. Jc3s5h (talk) 16:33, 10 February 2015 (UTC)

Probably the interface should be changed to reflect what you said, but for now it works like this, and it works in the intended way (right or wrong that it may be). In the future if the dev team decides to change it, they will probably have to change the values in the database. But until they do that, I don't see any other way out of this other than to insert the date as it this. Yes, it's convention-defying, and it's surely not the best way to do things. But in some way we MUST show the users the correct date. I don't think we can realistically expect anybody to understand that the date he reads he's not the "true" one and has first to be manually converted to make some sense. Candalua (talk) 17:38, 10 February 2015 (UTC)

In view of the devastating bugs related to this matter, I hereby interpret every date in Wikidata prior to March 1, 1923, as shit. I will no longer work on any such dates until the bugs are resolved. Do what you want, but whatever you do is wrong. Jc3s5h (talk) 18:12, 10 February 2015 (UTC)

Well, i wouldn't use that word ;) but yes, we must assume them as highly unreliable. Thats why I'm always trying to find a source for every date, so at least one can check the source. Please don't take it personally, I'm just trying to do the "less wrong" thing ;) , because i think that probably these bugs will not be resolved anytime soon, and in the meantime we have to go on in some way :( Candalua (talk) 19:08, 10 February 2015 (UTC)

I do not take your actions personally; you have been very reasonable. I'm not sure if one can add only a source without adding a date; maybe that would be the "least wrong" thing to do. And just because I have stopped working on any particular date, I will take part in the discussions to fix the bugs. Jc3s5h (talk) 19:35, 10 February 2015 (UTC)

I have removed the date of birth in part because Wikidata does not support the calendar of the Roman Republic, and Augustus's birth date, to the exact day, is not known in either of the supported calendars (Julian or Gregorian). Also, Wikidata's method of storing years before 1 AD is undergoing change, so even giving the birth date to the nearest year should wait until Wikidata is fixed.

• I've restored, but let "Gregorian" remain not only the data save model, but as data display model. May be it is not exactly correct, but since "Gregorian calendar = no change on display" it works for infoboxes. -- Vlsergey (talk) 18:46, 3 August 2015 (UTC)

Please stop edit warring and removing the date of birth. I added a date of birth with a low precision to at least something. Feel free to add or replace more data of births with whatever calendar model is your favorite, but don't edit war. Just add multiple. Multichill (talk) 17:14, 4 August 2015 (UTC)

Changed to 63 BC: dates from Wikidata are exported to ru-wiki, and "I century BC" as a birthdate looks very odd. By the way, determining the precise date of his birth is a complicated issue: Suetonius states that he was born under the Capricorn sign, and it's winter (Aug.94: "[Augustus] issued a silver coin stamped with the sign of the constellation Capricornus, under which he was born"). Homoatrox (talk) 08:24, 6 August 2015 (UTC)

I have found a modern source Bennett (2007) which flat out states:

Augustus was born on 23 September 63 B.C.(R).¹

Bennett's footnote 1 states:

Sutonious, Augustus 5. This article distinguishes dates in the Roman calendar from Julian dates by the notation (R).

I have removed the birth date 23 September 63 BC from the item on the strength of this source; since it is clearly a Roman calendar date, there is insufficient surviving historical information to convert it to the appropriate Wikidata-supported calendar, {Q|1985786}}.

Citation for the source, using Chicago style citation

Bennett, Chris. "The Two Egyptian Birth Days of Augustus." Zeitschrift Für Papyrologie Und Epigraphik 161 (2007): 195-98. http://www.jstor.org/stable/20191304.

Jc3s5h (talk) 17:45, 19 July 2019 (UTC)