The family tree is not a reliable place to post information for future generations

LegacyUser

Jim Kennedy said: I have been looking for a place to post historical family information and photos to make them available for future generations. The family tree on FamilySearch.com has features that could make it an idea place. However, the online discussions I found by people with a serious interest in genealogy were predominately negative. Many people have abandoned the FamilySearch family tree.

I decided to give the family tree a try to see whether the negative comments were from a biased vocal minority or actually represented common experience. My initial experience has quickly revealed that those who abandoned the tree as unusable for serious genealogical work were fully justified.

The time spent correcting the constant injection of unnecessary mistakes by people who do not know what they are doing appears to be greater than the time spent posting new information. Entering incorrect information is very easy and the policies allow and even encourage people to begin entering data with no training about good practices for genealogy or using the system. This is a formula for generating errors. The idea that the FamilySearch family tree will converge to increasingly accurate information over time is not plausible given the way the tree is currently managed.

These topics have been raised in the past, and should be continually raised until the implementation of the family tree is better aligned with the goals for the tree. One option would be to change the goals. The FamilySearch website could openly state that anything posted on the family tree is highly susceptible to erroneous changes and must be continuously watched. Users wanting a place to post reliable information for future generations long after the user is not able to watch over the website should find another website.

The FamilySearch family tree is currently managed as if it were a sandbox (isolated area for experiments and learning) for users with no experience in genealogy or using the system. If FamilySearch continues to implement this learning function in a way that sacrifices (a) accuracy, (b) long-term reliability, and (c) many experienced users, these priorities and trade-offs should be openly described to avoid false expectations.

An alternative option would be to have more balance between learning and accuracy that provides better learning and better accuracy. The family tree system could require that users go through a 15 to 30 minute online orientation module before being allowed to make changes to the tree, and require that sources or reasons for changes be entered before a change can be completed. These two requirements have been suggested previously by others and would not eliminate all problems, but they should greatly reduce the types of problems that are occurring now and that drive many people away. I expect that those resistant to such changes would soon discover that they do not want to return to the current practices.

LegacyUser

Cindy Hecker said: I am just a user like you, but Family Tree never claims to be perfect or a reliable place for record keeping forever. It tries to be a Free service, open to and including all people of all levels from child, to beginner to professional genealogist. A global Family Tree where we can include everyone with lots of tools, sources and hints to help us. It is open edit for a reason, and a place to work collaboratively. Have you read the overview of the tree here: https://www.familysearch.org/tree/ove...

When used correctly, we work together to make it better, when things are sourced and details added it will get more accurate. But there will always be new people we need to teach some things. There are many resources for that too. I have always been advised that you should keep a tree on your own computer using software to keep your information as yes this Family Search tree can change.

FamilySearch tree is not perfect and never will be but it helps SOO many! I am grateful to be able to use it and to help others as I work in the tree to make it better for everyone who comes along later.

Cindy

LegacyUser

Lynne VanWagenen said: If a strict preservation is what you're after, one way to do it is to build a family tree in an off-line product.

You can then submit a GEDCOM file to the Genealogies section. Other users can search and view the contributed files, but they can't change them.

https://www.familysearch.org/mytrees/trees: To Upload a GEDCOM
https://www.familysearch.org/search/family-trees: To search contributed GEDCOM files

LegacyUser

Paul said: It is only since I spread my net wider, to research more distant relatives (and some individuals not related to me), that I have discovered the truly shocking state of many of the branches in Family Tree.

When users ask for the ability to increase the number of individuals they can "follow" (from the current 4,000) I say to myself, "Are you sure you know what you would be letting yourself in for?" It was okay when I was just dealing with my immediate family, but now (from time to time) I can spend several days undoing incorrect merges and separating individuals from families to which they clearly do not belong.

Naturally, I have made mistakes and would not want to block people from using the program on the basis of their (lack of) experience. However, I am convinced that eventually FamilySearch managers will come to realise the present situation will become untenable if no action is taken at all. Perhaps that will mean greater "policing" - including sanctions (say, read-only access), training requirements, or whatever - but I cannot see the project ever being a success in its present, open-edit form.

For those who would disagree, I would suggest you might wish to share the pain of many of us, by picking individual IDs at random to investigate. I have done this with relatively common names (in England) like James Young. It almost defies belief how many careless individuals find it of so much importance to eliminate every possible duplicate that they merge (say) every James Young who lived in England and was born within a certain time period. Never mind that there are different spouses and children born in different places in the same years, the elimination of all possible duplicates appears to be almost an obsession with many users.

Collaboration and messaging? IF I get a reply to a "Would you kindly explain why you did this?" message, it is usually either a simple "Sorry if I made any mistakes" or "How do I put that right, please?" reply. I just advise them not to worry, be a little more careful in future and that I will sort everything out.

Of course, we have to back-up everything in software inaccessible to others, but what do we do about the errors in Family Tree? Ignore them, or just carry on spending many hours correcting them that could have been used so much more productively.

These arguments will not go away, so I feel it is time for there to be a serious debate and possible re-evaluation of the whole issue.

LegacyUser

Adrian Bruce said: To take up some of Paul's points, at the very least, the so-called collaboration tools need to be examined to see if they are being used for collaboration or not. For instance - are Discussions being used for genuine inter-user discussions? Or are they being used for notes simply because it's more difficult for anyone else to trash them?

Programmers are writing all these tools but (usually) no-one's telling the users of FS that they're there. The recent "Following" changes are an honourable exception to that but all too often I wonder if the techies are disillusioned with the take-up of new facilities. There would be serious consequences from a cost-benefit analysis of FS's software if this were a commercial environment - it's not but remember that people can only write one program at once....

It's also clear that the training materials are sadly deficient. Twice recently (and I think I'm lucky) I've had to unravel instances where one profile has been completely over-written with someone else's details. ("Invasion of the Body Snatchers" someone flippantly - but accurately - called it). In both cases the user appeared relatively new - I have really no idea how they managed to mess things up but it did appear to be an honest mistake. I can't help but think some better training on how stuff gets created would be better.

And as for "the name's the same" merging, yes, we can talk about building extra defences in the software - but surely the human brain should be capable of looking and deciding that profiles for Samuel X in Devon, Cheshire and Kent are probably different? (Especially since one required a time machine) But have these people even realised that the computer's hints might be wrong? Has any training told them that hints might be wrong?

Unless and until FS shows a degree of willingness in these directions, instead of simply counting index records and profiles being taken to ordinances, then any moderately competent genealogists will come to the conclusions that the whole edifice is built on shifting sands.

LegacyUser

Juli said: I read somewhere recently that Google Translate's crowdsourced "corrections" mostly have the effect of solidifying errors rather than actually improving anything. This goes exactly counter to the general belief out there that open-edit systems will eventually move toward a "fully correct" state.

The truth is, open-edit/crowdsourced will not, by itself, improve anything. For every correction someone makes in one place, there's a newbie (or outright saboteur) making errors someplace else. There have to be other sources injecting more-correct information, such as specific improvement projects, and there have to be strong tools in place for preventing the injection of less-correct information, such as levels of community oversight, accounts with training wheels for newbies, and account blocks and remedial training for those making too many errors.

Collaboration tools for improving the tree have to be primarily human-based: there's only so much the computer can do. No, the system shouldn't allow a Johann Jung in Vienna to be merged with a Johann Jung in Bavaria without an "are you sure?" message, but every now and then, such a correspondence may actually be correct. (Franz Liszt's maternal grandmother was born in Bavaria, married in one town in Lower Austria, and died in a different town in Austria. And her husband was sixty-something when they married.)

LegacyUser

Adrian Bruce said: I think that 2nd paragraph is particularly important. What is it that will raise the overall Genealogical Intelligence Quotient? Experience will only help individuals. Overall, the most experienced genies will, sooner or later, depart this mortal realm to be (statistically) replaced by newbies, leaving the GIQ unchanged.

What will persuade the newbie that the book that they have just discovered about their ancestor might not actually be unknown, might not be accurate, might even be fraudulent? Time, experience and a growing sense of cynicism will help. How do we accelerate people along that path? Juli has some good points there.

LegacyUser

Jim Kennedy said: What Adrian described as “body snatching” I think of as hijacking a person page. I have seen this several times in my limited experience with the family tree. Someone finds a person page with the name they are looking for and proceeds to add a spouse and children and change the birth and death information. The fact that the person page was well established with many sources that are inconsistent with the family they are adding is not considered. Such hijacking appears to be based on the assumption that there could be only one person with that name and/or that this is a personal family tree that should only have this particular user’s relatives.

The thoughtless merges that Paul described similarly appear to be based on the same assumption that there could be only one person with a certain name and/or that the tree should only have this user’s relatives.

Two basic points that any user should understand before being allowed to make changes to the family tree are that this is a shared family tree, not a personal tree, and that in the past there were different people with the same name. A simple orientation module that described these points and gave some examples could greatly reduce these relentlessly recurring problems.

As one who worked in systems analysis and software development prior to retirement, it is frankly inconceivable to me that users are allowed to make changes to the family tree without a basic understanding that this is a shared tree and that changes must be made carefully because different people had the same name.

LegacyUser

Jim Kennedy said: This is an interesting idea that I will think about for long-term preservation. However, my impression is that a GEDCOM file in the Genealogies section would probably not be looked at by many users, particularly those who are just beginning to search for family history information. I would also want the information in a more accessible place somewhere.

I have many hundreds of photos, obituaries, images of county records that are not online anywhere, and biographical summaries that integrate this information with newspaper articles and online census, marriage, death, and other records. I’m not sure I would trust a GEDCOM file to keep all this organized for several decades in the future.

LegacyUser

Jim Kennedy said: It appears we are making the same point. Those who want a place for reliable information to be kept for the long-term should look for another website.

As you point out, “when used correctly,” the tree works well. However, the tree is often not used correctly, and as you also point out, apparently is not intended to be a reliable place for long-term record keeping. I think that point should be made explicit on the website.

LegacyUser

A van Helsdingen said: I would support merges between profiles where the distance between the birthplaces and/or deathplaces exceed a certain limit (say 100km) could not proceed without a specific warning message being displayed to the user, e.g. "Are you sure? [Profile A] was [born/died] [km/miles] from [Profile B]. Ditto for time, maybe 10 years. Another possible idea is to let watchers of each profile object to the merge.

For even greater distances and time gaps, e.g. one profile born in Europe and the other in the Americas, or one profile born 100 years after the other, the merge should need to be validated by a FS Staff member.

I suppose one potential problem with the above is that some users will change the birth and death information on one of the profiles to be merged to be more similiar to the other, in order to avoid the restrictions on merging.

I'm fortunate that for many of my lines, I'm either the first or second person to enter them into FSFT, so my personal experience with poor merges and inaccurate information is fairly limited. But when I have come across these errors I know how time consuming they are to fix, and I sympathize with those who have to deal with this regularly.

LegacyUser

Don M Thomas said: Just another patron like yourself, and my own thinking.

In its present form, the "Family Tree" does not allow the saving of correct information on our ancestors, but it does allow for Temple work to be performed on our ancestors, and thus preventing the Lord from cursing the earth when he comes.

The "Latest Changes" is like a cesspool, collecting good and bad data about our ancestors. I am thinking that the most important part of the "Family Tree" is the "Latest Changes" or change log, and a lot of money should be spent on improving it. At a future time when there will be no deception on the earth, the "Latest Changes" along with "Genealogies," and home computer family tree management programs stored in "Genealogies," will be the starting point for all research.

LegacyUser

Jeff_Luke said: There are lots of great points being made here. I think there are a few things familyseach could do to help reduce the incorrect merges.

The rules around 'Possible duplicates' suggestions in 'Research Help' seem too broad. Familysearch seems to suggest possible duplicates that are often highly unlikely based on the records, or in some cases have literally nothing in common other than a similar name. Often the info is so limited that it would be very difficult to ever figure out how that person is related to anyone else. Less experienced users may trust those recommendations more than they should, or desire to 'clean things up' by merging.

For example, records of people named 'George Swartz' born in Pennsylvania, with no birth year, no records or sources and no other information. There have been countless George Swartz born in Pennsylvania - if there is no other info how could they possibly ever be merged with confidence? There are a lot of these 'orphans' floating around in familysearch that can never (well, should never) be able to be merged (as a side note, I wonder where all of these orphans came from and how they came to be in the familysearch system - is there any additional information that could be associated with them from some data source from which familysearch originally got the name?)

Another feature that may tempt people to do a merge without full confidence is that temple ordinances cannot be reserved when the system thinks there is a possible duplicate. That seems like a good rule to keep people from duplicating ordinances, but an unintended consequence is that some people may go ahead and make the merge in order to reserve the ordinance without properly vetting (again, since familysearch rules for identifying possible dupes seem fairly broad, a less experienced user may tend to trust those recommendations when they should not).

When familysearch does suggest a possible duplicate it would be nice if there was a way to categorize the duplicate as something other than 'not a match', but maybe as a 'possible match, needs more information' to allow ordinance reservation without forcing someone to make the determination that they are definitely a match or definitely not a match. I admit that is not a perfect solution, but maybe it would help someone think of some better ways to manage that situation.

LegacyUser

A van Helsdingen said: "Another feature that may tempt people to do a merge without full confidence is that temple ordinances cannot be reserved when the system thinks there is a possible duplicate"

With all due respect to the Church of Jesus Christ of Latter Day Saints (of which I am not a member), I am concerned by that. The current arrangements would be incentivizing bad merges. From a genealogical perspective that is not good, but surely allowing bad merges to create non-existent people who can then be proxy baptised is not good from a religious perspective as well.

I'm certainly not opposing this policy or attempting to interfere with internal matters of the LDS church, but simply pointing out that the negative side-effects of this policy must be mitigated. The ideal obviously is to prevent bad merges (which is negative from a genealogical perspective) as well as situations where genuine duplicates are not merged (negative from a religious perspective). A balance needs to be struck between excessively obstructing merges, leading to duplicates, and not preventing bad merges. The current arrangements are closer to the latter.

LegacyUser

Adrian Bruce said: Re "Another feature that may tempt people to do a merge without full confidence is that temple ordinances cannot be reserved when the system thinks there is a possible duplicate"

And AvH's reply:
"With all due respect to the Church of Jesus Christ of Latter Day Saints (of which I am not a member), I am concerned by that. The current arrangements would be incentivizing bad merges."

I am also not a member. As AvH's reply indicates, a balancing task is needed. I don't have any issue with checking for possible duplicates - that's the last phase I do when entering or checking profiles in FamilyTree, and I would agree that such a check should indeed be mandatory. The unfortunate result, I think we all agree, is when people "believe" the possible duplicate message and clear it up by merging, even when the 2 profiles were not, in history, the same person. I am, however, unclear how effective that overall processof checking for possible duplicates actually is. In Rotkapchen's saga of duplicates created by loading of GEDCOM files, from memory, there are plenty of instances where duplicates have been created and set on their way to ordinances even though the dupes have already been through that process.

So I think I would say that it is difficult to say where the perverse incentivizations lie - not, I suggest with the checks for dupes highlighted. If there are perverse incentivizations, they lie with the desire of Church members to be progressing the Church's "recruitment" work in the after-life (apologies if that phrase upsets anyone) at the possible expense of genealogical accuracy. It was a Church member, IIRC, who described younger members being upset because they couldn't find any relatives to set off on the temple processes because they'd all been done already. Well, if your families have been in the Church for generations, in the middle of Utah (say), it stands to reason (it appears to me, rightly or wrongly) that there's a good chance that your low hanging fruit have already been taken.

I don't have a problem with the Church and its interest in genealogy - I believe it can be a win-win for everyone. But I do think the Church and FamilySearch need to ask serious questions about where the emphasis on quality v. quantity should lie, particularly for families of long-standing, and review the strategies in that light.

(It would be nice to say both quality and quantity apply. Saying it is easier than facilitating it!)

LegacyUser

Paul said: I do hope FamilySearch employees (including "management") are following this thread and accepting the comments as genuine concerns, coming from Church members and those who are not. To me, this is constructive criticism of the best kind and should give rise to give serious thought on how Family Tree might be improved, without necessarily changing the general ethos.

Most of the points raised here are probably not new ones. In which case, I will raise a point that I have put forward in an earlier thread:

What is more important, urgently eliminating as many duplicates as possible, or trying to ensure merges are avoided that will lead to individuals being treated as if they never existed on this earth?

We have been promised so many times that the "possible duplicate" algorithm, which leads to "FamilySearch" adding these suggestions to the person pages, will be made far less loose. Yet still I find highly unlikely ones appearing. The enhanced feature for assessing / making merges SHOULD make it a lot less likely for an incorrect merge to take place, but still many users continue to ignore the fact they are only dealing with "dupe" suggestions, and seem to make a merge every time one appears.

I don't envy the developers and managers who have the task of trying to strike the various balances. In theory, there should be no need to "trade off" one factor against another, but even if my / our ideas are not quite on the mark, the status quo certainly does not seem to be a viable option.

LegacyUser

Don M Thomas said: Comments leaning to the negative?, who would have thought in years past. Shows what an intelligent Feedback can do.

It appears that FamilySearch is a collect all of ancestral data, (good and bad). More money to promote ideas and technological innovations should be spent on the "Latest Changes" or change log.

Home computer family tree management genealogical databases should be uploaded to FamilySearch's "Genealogies." Click on "Search" and then "Genealogies" above. Then click on "Submit Your Tree."

There are the Granite Mountain Records Vault which is thinking towards the future.
https://en.wikipedia.org/wiki/Granite...
https://www.bing.com/images/search?q=...
https://newsroom.churchofjesuschrist....
https://www.bing.com/videos/search?q=...
https://www.bing.com/videos/search?q=...

LegacyUser

Jim Kennedy said: The comments in this thread have shifted more to suggestions for consistency checks, warning messages, and algorithm changes that could improve the family tree. Those are important ideas that are interconnected with the original questions about the purposes and goals for the family tree. Here is a summary of my current understanding about the purposes for the family tree.

Given the discussion here and the way the family tree is being managed, my impression now is that the managers of the family tree have made firm decisions about the purposes of the tree, but those decisions have not been clearly and transparently communicated to users. That lowers the credibility of the website and the sponsoring organization. I suggest that the description of the family tree include in a conspicuous place something like the following to introduce much needed clarity and transparency about the purposes of the tree:

“The FamilySearch shared family tree has the primary purposes of social networking about family history and genealogy, providing a place for people to learn about genealogy, and implementing the ordinances of the Church of Jesus Christ of Latter-day Saints. Any information in the family tree is subject to change by others, and the changes may be correct or incorrect. With constant vigilance and efforts to correct errors the family tree or parts of the tree may become increasingly accurate over time, but that cannot be assumed given the frequent mistakes by those in the early stages of learning about genealogy. This shared family tree is not intended to be a reliable place for long-term retention of family information.”

If the actual purposes and priorities for the family tree are openly described, some people will probably stop using the family tree. But, most of those people will probably leave anyway. The question is how much time will they waste and frustration will they experience while they learn what the actual purposes and priorities are in practice. I am a case in point.

LegacyUser

ATP said: Jim Kennedy, your observation and the summary of your understanding about the purposes of FSFT is completely on target.

And, the statement regarding its credibility and its impact on its sponsoring organization could not have been better said!

Further, your purposed introduction to the user could not be clearer!

And, your final paragraph sums up the possibilities, yeah, probabilities of what happens when a one encounters any project that is so far removed from its organic purpose, they will leave.

As a member of the Church, for awhile, now, I've been trying to figure out how reconcile the commitment to my kindred dead - which is you may know, Church talk for temple ordinances - and remain a part of FSFT, knowing in its current rendering, that it is not a reliable place for long-term retention of the work and effort I've put into it for the past year and a half. That work amounts to some 15,000 changes and adding more than 12 and half thousand Sources! Most of those sources were original to the paper pedigree charts and family group sheets submitted decades ago and disappeared into the memory hole, somewhere along the way, along with the temple ordinances.

Thank you, for the clear and explicit articulation and for your desire to share your knowledge and understanding of genealogy and family history research and its transfer from analog to digital in an effort to create order out of chaos of this site.