Gramps is great for logging exact genealogical data. But how do you log things with mild ambiguity?
For example: Places. I have some Place names that are ambiguous. There are municipalities within counties of the same name. And municipalities that straddle county or state line. And popular place names that were used for several distinctly different places… but within a 10 mile radius in our tri-county area.
Or… for Persons having ambiguous relationships. DNA tells that someone is a maternal or paternal cousin of some degree of separation. (Although the distance may be greater than expected. Pedigree collapse could mean a double cousin appears to be more close than actuality. Or random genetic chance leads to a larger than normal correlation.) But the actual common ancestry is unknown.
So how do you log these People & Place so the Gramps can help indicate the correlation?
I have my methods that I will share later. But I find them awkward and so wouldn’t want those to stifle alternative suggestions.
I will digress from the questions but not completely.
First of all, I don’t do my DNA genealogy but if I did I wonder if I would not duplicate my database, one regular genealogy DB and one DNA genealogy DB, in order to represent these two different things which are the “paper” families and the “natural” - biological - ones. Suddenly the ambiguity mentioned would arise less. The question could then be how to reconcile the two genealogies (if it proves necessary?) in case of divergence. In this case it could be interesting to publish these two genealogies and linking people and families with web links (the best would be if we could link identical people between two different DB). Someone suggested that I do my “spiritual” genealogy, linking children to their godparents rather than their parents (to do that I would also dupplicate my DB). Doing its DNA genealogy looks a bit like this in my opinion (except that between the usual genealogy and the DNA one we can assume that there are more matches than mismatches than between usual and godparents genealogies (?))
Concerning places, why, as for persons (personaes), not duplicate them by associating them with a common note to do in order to seek to resolve the ambiguity or at least warn readers that it exists? An object is an object, especially in Gramps and python, why treat them
conceptually differently.
In both cases, “objects or classes” differs, so why not dupplicates them.
Your problem with places sounds very unique to the US and countries who made their subdivision borders with a ruler on a paper map, dividing existing villages in the process of creating straight border lines.
I think most of the time places with the same names can still be distinguishable through the place hierarchy.
More tricky are those places if they have the same hierarchy e.g. the same county. I’m unsure how to solve this, but I wonder why they weren’t renamed to be more distinguishable.
In my DB I have the problem that some places have several roles in the subdivisions of a country e.g. Vienna, the capital of Austria, being a federal state, political district and a municipal. I decided to create one object for each role, because this way it is easier to select all places of one subdevision level which I need to do often. A disadvantage of this is that the place hierarchy looks more confusing and all objects show up in the place selector.
Persons
I use tags and research notes for persons where I suspect a relationship, but wasn’t able to find it yet.
For DNA I do add the persons I match with (only those with a high cM and high chance to solve), link them through associations and tag the persons. Other matches are just saved in a list.
When I’ve found the connection through persons, I’ll add them then. If I’m unable to find one I’ll keep the known family trees seperated, except for the people I’m interested in e.g. those who lived in the same area of my ancestors (neighbors).
I’d try to avoid adding empty placeholder people and families as links, just because a person is probably a x-degree cousin.
I’m not aware of any that straddle state lines. Maybe you are thinking of Kansas City, Kansas and Kansas City, Missouri (for example) but those are two different cities, each deserving its own Place.
However, cities do sometimes straddle township and/or county boundaries (as in this previous discussion). In those cases, any particular address within the city would be in only one township and county, but it’s problematic in the place hierarchy.
Still somewhat on topic, I think, addresses can be ambiguous too. Streets are sometimes renamed, or the houses renumbered, and houses themselves are sometimes picked up and moved to a different address (this happened to my relatives’ house). But I don’t know how to deal with these things, other than adding Notes.
The usual problem is that no address is included with the record. Only the ambiguous place.
Since addresses & places can be date limited, then those particular ambiguities can be addressed directly. Better handling of changing geopolitical boundaries enclosing a Place is usually mentioned as one of the Gramps 5.2 targets.
About this particular point, at least three France cities - Paris, Lyon and Marseilles (why a final s in English?) - are divided into arrondissements (districts). So streets can be in one arrondissement or can cross more than one arrondissement. Usually the hierarchy is: street number, street, arrondissement, city. But in Gramps it would be more correct to use that hierarchy: street number, arrondissement, street, city. But it’s not pretty (and multiply arrondissements while v5.2 new hierarchy isn’t there, here for place and administrative hierarchies), so I don’t use it and I know I’ve some volunteer errors in my database.
I have lots of ambiguity, or perhaps “apparent ambiguity,” that I wish to capture when I run across it if for no other reason to know I’ve run across it before and should not concern myself with it when I run across it again. Unless I want to dig into it, in which case I’d like to have everything I dug up before handy.
Sometimes the ambiguity is resolved negatively - “Nope, that James Parks is not connected to the James Parks family of interest, at least as far back as the person who arrived in America in 1817.” But is that resolved? They could be cousins via a relative in Ireland before either made the trip to “Amerikay.” What if I find the common ancestor?
I am straddling two solutions:
I put my research notes and any documents or DNA information into Evernote, and work from there until I see a relationship, at which time I add the person(s) to my Gramps tree.
(Note: I have set Evernote to automatically import any media I add to my media tree, which is in the local copy of my Google Sync tree. This way, when I toss in a bit of media, both Google and Evernote make separate backups on the web. Evernote has the additional benefit of allowing me to add notes to the note where the media is living, and search on those notes AND any OCRed text in the document. )
I add the people to my Gramps db and do not connect that person or tree fragment to the main tree, but I make a note about the possible relationship (if DNA) or lack of relationship. Then, as I research a tree member, I will for sure find people with the exact same name, sometimes with the same birth year, that are not paper relatives. I find it helpful to track these people in Gramps knowing they are not paper relatives. When I run across them again, I can use the information to help determine if the new data belongs to a relative or the “other.”
The benefit of having non-relative people in my Gramps db to me is they show up when I search and remind me that this new information may belong to a non-relative, look closer. I don’t have the discipline to always check in Evernote for an alternative person before happily adding bogus data to a relative.
So, given the above, I am happy to enter people just cuz I have some DNA connection to them and leave them hanging with a note or tag that lets me know they are here because of a DNA test. I add DNA to the confidence tags.
A place hierarchy of street number, street, arrondissement, city. should work fine. You would just have the same street names in the different arrondissements.
Think of it as you are giving a name to the “spot”. For the most part the place database is never seen. It is only used by us, the Gramps user, to create the place name seen by our relatives in reports, etc.
