Introducing the Geni Consistency Checker

Mike,

Using the last name for birth surname will be the best fix for those profiles who where entered for man when the label for surname at birth was maiden name. Most of those profiles still have an empty surname at birth and a significant part of them the value of last name is a better value than that of the father.
I will see if I can find you some examples.

When I first saw this new feature, there were about 73 inconsistencies of various types showing up. That figure has since been brought down to 5. However, in 2 categories I have issues. Under "Children with different last names" I have ignored for all users 11 out of the 13 remaining instances, as the data are accurate. The 11 remain hidden, but not removed from the list. Under "Too old parent" I have confirmed from his own daughter that my great uncle had not one, but two children from the age of 75. This has also been ignored for all users, but remains hidden, not removed.

Are there any plans to permanently remove "ignored for all users" entries once they have been verified as accurate by the user. Let's face it. Even if the user turns out to be wrong in the odd instance, this error can always be corrected.

I look forward to your feedback.

Private User I believe we've got the hidden count and duplicate rows fixed. Let me know if you see any more weirdness.

Job Waterreus it should not be a problem for a male profile to have his surname at birth in the last name field. This is why I need examples, so I can see where this is going wrong.

Gervais de Matas when you ignore an inconsistency it's removed from the user interface everywhere except the hidden count / "show all" link in the header. We keep track of *when* an inconsistency was marked correct so we do have the ability to pass it over even in the hidden count / "show all" interface, maybe after it's been marked correct for a certain period of time. I'm not sure about the implications for that, though; we'll have to give some thought to what circumstances should revive such inconsistencies.

Mike Stangel - Please don't report an age based on a "circa" date as if the date was exact

1) one of my inconsistencies: "married when she was only 14 years old" -- how can Geni claim that when she has a birth year marked as "Circa"? - at the very least, the message for the inconsistency should say "married when she was only about 14 years old" - and considering that "17" is also "about 14" - not sure at what point it should be flagged as an inconsistency - but at the very least - not report it as exact if it is circa
2) This might be a good - or at least helpful - time to change display of dates on Timeline Events to not show as an exact date / exact year if the entry is circa (ditto for before, after, and between if those are still showing as if exactly one of them)

Private User example of the 14/17/21 year old please

Mike Stangel,

I saw some examples of what Lois is reporting, I will see if I can also find you some examples. You indicated that for the inconsistencies you would change it to handle circa differently for when only a year is given and when a full date is given. Could it be that this is still from the older check?

I noticed just now that changing the marriage date does not trigger a recalculation for an estimated birth date (when no birth date is given) and so the inconsistency for too old does not change when based on a birth estimate.

Example where I changed the marriage date which was 12 to 12-03-1729 (and removed the christening events with dates 25) without triggering a recalculation of the estimated birth (was between 2 and 2): Marritje Jans Portengen
I first added a baptize date 25-10-1696 and that also did not trigger a recalculation, so I also added an estimated birth date and that finally triggered the removal of the inconsistency.

Mike Stangel - sent you a PM with link to the profile with a birth year as "circa" but where I was told she married "when she was only 14 years old" -- not "about 14"

Two separate issues:
1) It should not phrase the message for "circa" as if it is "exact"
2) at what point, if birth and/or marriage is only circa, should it warn about too young a marriage

Lois,

I had that issue with Mike before (and I still agree with you and disagree with him).
It is okay to trigger the inconsistency, because there could be a problem, but the text of the message should be different.
Example: https://www.geni.com/inconsistencies/for_profile/6000000085310728134
The message could be from:
Beatrix Rietbroek married when she was only 13 years old.
to:
Beatrix Rietbroek could have married when she was only 13 years old.
That should be clear enough with the tip which mentions the before date
Tip: Correct the birth (born before 1754) or marriage date or marriage status of the profile (Beatrix Rietbroek).

I think it also would be clearer if the lifespan next to the picture on the inconsistency page (like https://www.geni.com/inconsistencies/for_profile/6000000085310728134) also had indication of circa and before/after

Mike,

In answer on https://www.geni.com/discussions/203863?msg=1343705
The problem is not so much with the surname at birth being in the last name field.
The problem is the surname at birth is empty and filling it with anything else than the last name (like the last name of the father) will often put wrong data in it.

One (not so good) example would be: https://www.geni.com/inconsistencies/for_profile/6000000038168958266 (the inconsistency for missing surname at birth)
In case like this (mostly for profiles that where entered when surname at birth was labeled maiden name) the surname at birth was put in the last name field.
When you use Caan you get different values for surname at birth and lastname.
In this case it would be better to have Kaan in the surname at birth and Caan in the also known as.

Suffix in name still triggers on last names (like de Ridder) where there is a preposition not starting with uppercase in front of the name. Example: https://www.geni.com/inconsistencies/for_profile/6000000001958115961

Job

Thanks. Hope we can eventually convince him -- Anything other than "exact" should not be treated or phrased or shown as if that date was a date entered as an exact date.

So far, have been going from my 'Full' inconsistency page, so had not noticed that Lifespan that is on the Proifle inconsistency page - agree, that totally should be changed also!

Today I found out that most of my (female) profiles are marked with the "missing birth surname" inconsistency. However, most of my family follows the Spanish naming customs, so there is no such thiing as birth surname, just surname.

Moreover, geni suggests using either the father of the mother´s surname, while it actually should be the mix of the two (first last name of the father, first last name of the mother).

Is there any way to indicate the last name format (e.g. Spanish), so we don't have this problem?

For Dutch titles these should also be added if they are not already on the list:
Male: koning,prins,hertog,markies,graaf,burggraaf,baron,ridder
Female: koningin,prinses,hertogin,markiezin,gravin,burggravin,barones,jonkvrouw

Private User,

There is a setting https://www.geni.com/account_settings/name_preferences but that is really meant only for displaying a name and does not have a setting for this.

I think Mike Stangel should answer this.

The inconsistency for missing surname at birth does not seem to recognize Dutch patronymics. Example: Jouck Dircks

Most inconsistent for users is following a false system, indicating long lists of people named 'van/van der/van de /von/de/de la/du etc followed by tha centername like my name 'Oeveren'. I think 80% of my family is categorized in the Index-letter 'V'. I have to scroll till I finally drop dead myself ! Some people are creative by filling out the name as 'Van' Oeveren' in order to show the name at the beginnging of V. Where a lot of noble people from Germany(von Aesch,von der Aesch) France(de La Palma, du Bois) in a family are stated in de tree-index it is really bizar. And finally it's slowing down the system dramatically ! A father named 'Van Oeveren' with a son 'van Oeveren' and a daughter 'n.n'./N.N. or van/Van Oeveren the names are scattered all in de (index) letter 'V'. - User are often mistaken by these facts !

Thanks Job Waterreus

I think changes should be much deeper. For instance, that the user could select a surname pattern (Anglo/Spanish/patronymic/etc.) for each person, so it automatically suggests the last name of the father, father/mother, etc.

Even more, it would be awesome if we could link the origin of each surname or surname particle. Like, for each last name, indicate if it comes from the father, mother, husband, etc. This is specially relevant for older records, where name customs were much more inconsistent.

Right now Geni it's trying to force the Anglo name convention to all profiles by default, asking for a "maiden name" and a "married name" (but just for women, apparently), which in my case I have no use for.

Another consequence of that is that when people have two (or more) last names (by default in the Spanish and Portuguese-speaking world), Geni treats them as just one, making the Index of Last names unusable, because it treats every single combination of last names as a separate one. (e.g. "García González" as just one last name, instead of "García" and "González" separately).

Mike Stangel,

Would you also include suffixes like "d.e." and "d.y." These are an abbreviation of "den eldste" (the oldest) and "den yngste" (the youngest) in Norwegian. In Norway, these are common to use in family books and genealogy to differentiate between siblings with the same name. They were also (sometimes) used during the period when these people were still alive and not only in posterity. I have seen them being used in old church books.

It was common in Norwegian family history to give the next child in a group of siblings the same name as the previous one if it died before the next one was born. The same is true if a married couple had two parents with the same first name, then it was not uncommon that two of the children could get the same name. Naming traditions was strictly followed. Occasionally, the naming of grandparents with the same name was given to only one child, but sometimes the parents decided to "honor" both grandparents by naming two of the children after each of these, even though they had the same name. And many of these kids would also grow up. It is therefore easy to make mistakes of these and believe they are the same person even though they wasn't and had a separate life from each other as adults.

If the child mortality in a family was large, it was also common to have more than two children of the same name. For example, if the son "Olav" died as a an infant or small child, the next son would be called "Olav", but if this child died before the third son was born, this one would also be named "Olav" (for the third time). This way it continued to a higher number. I have seen several examples of this in genealogy. Therefore, in some cases, the suffixes "d.m." which is abbreviated to "den mellomste" (the middle [one]) should be used. If there are even more people with the same name, you must use "d.n.y." ("den nest yngste", the second youngest) and "d.n.e." ("den nest eldste", the second oldest).

Thus, the suffixes "d.e.", "d.n.e.", "d.m.","d.n.y." and "d.y." should be added to the system with "lower case", "period mark(s)" and "no space" between. As listed in order: the oldest, the second oldest, the middle, the second youngest, and the youngest. They are used in the same way as "Jr." and "Sr." is used in English-speaking countries to differentiate between father and son with the same name (or mother and daughter).

Private User,

You could add each separate name to the also know as. That would make them searchable. But still it needs a better solution.

Mike Stangel,

Another example for where using the last name for surname at birth would be a useful fix: Imke Sjerpsz van Dijk

adding to the Dutch titles Job Waterreus listed earlier
male nobility: jonkheer (can be abbreviated jhr)
academic titles: prof, dr, ir, ing, drs, ds, mr (all in fact abbreviations and all both male and female although some feminists replace drs by dra)
and mgr (catholic clergy only)

Private User,

The Dutch naming system with a prefix like "van", "de" or "van der" in front of the last name starting with a lowercase letter is rather unique. If you look at Belgium names you will see that they include it in the last name and start it with an uppercase letter and sort it accordingly.

If may be nice to have some inconsistencies especially for the Dutch name conventions (but lets first try to get the current ones to perform a little better).

Mike,

Still another example where using the last name for surname at birth may be a better solution than the name of the father: Jacob Jansz Broers

The current inconsistencies have a lot to do with that. When i.e. the Fathersname is different in a minor way, a consistency pops-up ! Father 'van de Lei & offspring van 'der' Lei' . Furthermore the ancient books stating 'Vande' Lei etc. It's the same in France with surnames "de la Palma'. I think that that the majority of problems more have to do with the prefix-surnames and should be taken action for an extra FIELD named PREFIX OF SURNAME. When the actual problems are surpressed, no action will be taken to make the system better and faster! It happens too often that "GENI will be right back" !

René,

The inconsistency will also check the also known as field.
Some of the inconsistencies can be avoided by putting all the spelling- and name variations in that also known as field (use comma's in between as separator)

When you know there is no real problem (people do change their names) you can ignore the inconsistency.

Mike,

Could the name of the father be added to the inconsistency for no birth surname.
That would make it much easier to spot if patronymics are used and the inconsistency can be ignored.