Re: [Gramps-devel] Minor DATE problems in GEDCOM-export

[Lars Kr.L. <gr...@lk...>]

On Mon, Dec 09, 2002 at 09:17:46PM -0700, Don Allingham wrote:
> I'll see if I can put together some rules for dates. Before I implement
> them, I'll post them for commnents.
> 
> Don

Thanks.

In that case I will make a comment on rules regarding the
changeover from the Julian to the Gregorian calendar. :-)

Dozens of programs generate invalid GEDCOM-files (and have done
so for years) and I think it would be great if an open-source
project like GRAMPS already at a very young age could raise
itself above that.

I have my own genealogy in a bunch of LaTeX-files and I generate my
GEDCOM-data from these files with a perl-program I wrote. (This
approach may seem awkward but it has evolved from my start in
'89 when there weren't a lot of genealogy software to choose from.
I now import my GEDCOM into gramps and play with the data there).

As my research has taken me back through time I have ancestors
in places that have different dates for the changeover to the
Gregorian calendar.

Knowing about changeover dates is necessary to properly check for
invalid dates (and to properly set the DATE_CALENDAR_ESCAPE in
GEDCOM-export).

As far as I know there are about two dozens different changeover
dates used in various places. So I believe that robustly detecting
the changeover date from the place-name is not feasible.

In my LaTeX/perl-setup I deal with the different changeover dates
in this way:
For each changeover date I have defined a variable to indicate
usage of this changeover date.

If a person in my genealogy has such a variable set it means that
any dates associated with this person uses the according changeover
date. With no variable set the default changeover date is used.
If a person moved from a region with Julian calendar to a region
with Gregorian (or vice versa) the effect of the variable can be
overruled by specifying the calendar used for the specific dates
in these exceptional cases.

The default changeover date is that set for the default/home person.

A very special situation exists in Sweden from 1700 to 1712
(there must be many people in North-America with Swedish ancestors):

If one finds in a Swedish record of f.ex. burials a date like
12 SEP 1709, then this date is one day ahead of the Julian
calendar (because the Swedish skipped the leap-day in 1700 and
had an extra one in 1712).
The GEDCOM-standard does not deal with this so instead I put
in my GEDCOM:
1 BURI
2 DATE @#DJULIAN@ 12 SEP 1709
2 NOTE Swedish calendar: one day ahead of the Julian calendar

Perhaps some of the above could be useful in GRAMPS.

There is an article about various date-issues in genealogy on
http://www.genfair.com/dates.htm

Thank you,
-Lars Lundin
--
GEDCOMP: An extensive and free database for genealogists with
interest in Denmark: http://www.lklundin.dk/gedcomp/