Then Person, Family, Source, Repository, and Place would all be subjects.
It would probably be beneficial to then add a boolean so if a source is an artifact or heirloom it can be easily identified.
This then would do what you did not want to do. But a research subject may be a person, place or thing. A Source is a thing, as is a Group.
I think the issue with using Source for heirlooms is that it is an abstract concept in Gramps. It is the repository reference that identifies the physical instantiation of the source and a source can have more than one of those, and thus be a microfilm in one repository or a book in another. I see this comes from Gedcom allowing 0:M so is a flaw in their model. Media type should be on the Source record. A copy of a book may exist in multiple repositories. A microfilm of that book has a different provenance and is thus a different source.
A Repository is really a Place and should not be a separate object.
We can investigate the idea. What happens if Person, Family, Source, Place, and in the future, Group are subjects?
Suppose you want to record the purchase of a family bible. The event would have one person with the role of buyer, another with the role of seller, and the bible would be a source also pointing to the event. The bible would also be available as a source.
In a similar way, you have an event with a ship (a place) as a subject, leaving a port (another place). The ship would still be available as a place, which would be useful for recording events that happened on board a ship, but its exact location was unknown.
@Nick-Hall@cdhorn Could you please create GEPS for the “group” & “subjects” ideas, so there is an overview? Both new concepts are interesting with a lot of potential. Discourse is great for discussing, but it’s hard to keep track on what was agreed and what was scrapped.
What about the possibility of extending Associations beyond the Person object? Instead of Roles, the Events could have a ‘Reciprocal’ associations to each other. If the Sync Associations functionality was integrated as an option in the Edit Associations, then there could be a true relationship defined.
Gramps could Associate Emmigration and Immigration Events in the same fashion, setting the endpoints of the trip. If Associations were not required to be sequential, associating a Waypoint event with an emigration OR immigration, the trip appropriation could be expanded from a simple straight line to be a multi-segment ‘curve’
@Nick-Hall the thought occurs, given @ennoborg comments on the Research Journal and my sitting and looking at DeadEnds again, that Subject can have a SubjectRef list and we would have what we need to support evidence/personas but for any Subject.
Ideally a Subject object would encapsulate AttributeBase, CitationBase, MediaBase, and NoteBase. It would add a subject_ref_list, event_ref_list, and perhaps a boolean conclusion indicator.
Place lacks AttributeBase though, but if a ship or house/building is considered a Place then it needs to have them anyway.
Source uses SrcAttributeBase and IndirectCitationBase. To treat Sources as legitimate subjects we would convert them to their normal representations. This needs to be done regardless if they are to be used to record heirlooms and artifacts.
By treating an Event as a Subject you automatically have an event hierarchy as well.
That might seem like it is overcomplicating things but that really seems to make sense.
Once you add Group, and some kind of Source heirarchy for representing catalogs, and Entity as a top level object you basically arrive at having implemented a large portion of the DeadEnds model.
Yes. Place attributes and place events are part of PR #809 and GEPS 45.
Breaking this PR into smaller parts would be helpful anyway. At the moment it is too large to easily review. It’s unlikely to get approved without changes - I’m still unhappy with the place type changes. However, we could merge parts of it.
@prculley I don’t want to duplicate work you already did. Do you have plans on trying to split the Place Enhancements PR up into a few smaller parts?
And then @DavidMStraub you are looking at Gedcom 7 imports right? To fully support those won’t we also need to update many of the objects?
Wouldn’t it make sense to try to figure out what changes need to be made to all the objects, and make all of them at once instead of having multiple PRs changing things and touching much of the same code?
Spending some time reviewing that and revisiting the standard yet again but more carefully. I think some bad choices were made but that is just my opinion.
The question I am now asking myself is that it is nearly 17 months since the standard was released. That is more than enough time for any of the large genealogy players who certainly have the financial resources with dedicated development staff, some of whom participated in the development of it, to implement and provide support for it. Is anyone aware of anyone who has?
I do see Ancestry updated their Gedcom exporter in July 2021 from 5.5 to 5.5.1 and then updated it again in August of this year but it remains at 5.5.1 So that would seem to indicate they have choosen to refine and stick with that though perhaps they are just taking their time.