Bug still in play. Living people load as deceased.

karenrasmussen

I uploaded a new GEDCOM today. Title: CARL CET Tree. GEDCOM selections were all in file, names only on living, notes and sources. No other extras. However, loaded as a CET tree, living people show as deceased.

mynameistk

@karenrasmussen thank you so much for your help with this bug. What is the name of the latest CET that you uploaded?

Robert Kehrer

@karenrasmussen,

Karen, I have asked our software engineer to triage this specific incident, we accessed your uploaded Gedcom and compared the data in the file to the tree persons created in the CET. Here’s a summary of findings. 

1) A need to make it more clear how the system works.

The FamilySearch system requires that a tree person be labeled as either living or deceased. It is not an option to leave this undesignated. There is no Gedcom way to specifically designate a person in the file as “Living” separate from the conclusionary or evidentiary data. Our system determines living status based on the following rules:

a) Does the person have any death, cremation, burial or interment information? (If Yes= marked as deceased)

b) Does the person have any life events or data within the last 110yrs? (If no -> next step)

c) Do any of the surrounding relationships (parents, spouse, children, siblings) have any life events or data within the last 110yrs? (If no -> mark as deceased)

The take away is that any person without any kind of death info, nor any life data within 100yrs, nor relatives having data which would indicate the person could be <110yrs old, are marked Living. The absence of any such data gets them marked deceased.

2) A possible need to modify how the system works

a) We have seen a few persons that have the word “Living” in the death field. It’s rare, but we will be marking these as Living even in the absence of other data

b) We are considering examining additional generations for data that could indicate the person was born <110yrs ago (Grandparents, etc) 

c) We are considering whether we should be marking persons where they and their surrounding relatives have no data as living instead of deceased.

I’d value your thoughts, especially regarding the possible changes. 

Robert Kehrer

Here's an example from your Gedcom. Expanding to look at the grandparents (Arne Poul Rasmussen & Ellen Peterson Scholcoff), for example, would have informed the system that their child (Carl) and all Carl's children should have been marked as Living.

The two changes: 1) no data = Living & 2) Look for data on Grandparents, are two code fixes that we expect to have deployed for all future uploads within a couple days.

I will respond here when the changes are pushed to production.

In spite of the original bug, exposing living data is truly a serious issue here at FamilySearch. I've spent a good amount of time with our InfoSecurity team making sure that the original bug is understood and steps are taken to ensure the bug doesn't happen again. The software routines are pretty robust to ensure that, once a person is identified as Living, only the tree owner and invited tree members can see it. Unfortunately, due to Gedcom not having any way to designate Living except in the data, eliminating all the data on the living except the name, eliminated the current algorithms ability to accurately determine living status. Defaulting to Living status in the absence of data may generate a lot of false positive living tree people, but it is the safest path forward.

mynameistk

@Robert Kehrer I think option 2 - if not specifically marked as deceased, leave them as living. Sometimes birthdates are so wrong that you can't assume anything using them.

Susie O'Neil

I agree that people with no dates should always be marked living upon upload.

I just uploaded my first tree to test, which is very barebones with most people only having a birth and/or death year. As the tree is based on connecting my DNA matches who I don't know (or their living relatives), there are many people with no dates and they are being marked deceased. I had to make the tree private for now, as it is a lot of people to fix.

Robert Kehrer

Susie, no promises, but we are hoping to have the changes discussed above deployed early next week. I’ll post here when that happens. It will only affect newly uploaded trees after the code deploy, not previously uploaded trees.

MrsLCJ

I think that, if it is possible, then persons in a CET tree who are born before 1899 (>126 years ago) and their predecessor ancestors, should be marked as deceased. There may occasionally be persons aged 126. They'll be rare.

mynameistk

@MrsLCJ Thank you for testing early access CET and providing feedback.

mynameistk

@Susie O'Neil Is your CET now functioning as expected? tkp

mynameistk

@karenrasmussen @Susie O'Neil

Thank you for bringing this to our attention. We’re committed to resolving as many issues as possible, but unfortunately, the report did not include enough information for us to investigate further. Since no additional details were provided, we’re closing this discussion for now. If you or others encounter this issue again, please start a new discussion and include relevant information such as your operating system, browser, URLs, and screenshots if available.

We appreciate your support and engagement. Please don’t hesitate to reach out with any future questions or concerns.

Thank you for using CET.