Merging Duplicates - third option
When merging duplicates, there are two options: merge or not. I suggest a third option. When there is not enough info to be sure - mark it is "unsure" or "not enough info". That way it is not thrown out when it may indeed be a duplicate and it will not be merged at this time. Hopefully more info will be added someday that will allow a concise decision to be made.
Comments
-
This has been suggested on a number of occasions, over a period of many years. I think its lack of implementation is connected with there being nowhere for the "unsures" to go. If they remain on the page, they would possibly be seen as causing clutter; if they disappeared from the page, their status would effectively be the same as if you had selected "not a match".
Are you suggesting to the developers that these hints should remain attached with a clear label, to warn other users (and yourself, for a later time) of further investigation being necessary before making another (or "final") decision on the likelihood of a match?
2 -
@W Andrews , your suggestion could go a long way toward preventing incorrect merges, which are so prevalent in Family Tree.
@Paul W , you raise a good point about where "unsure" duplicate hints could go. What about on the Dismissed Hints tab, but they could be labeled "insufficient information" or something like that, instead of "not a match." That way, they would not show up as possible duplicates any more, but it would be fairly simple for users to view them.
Also, if someone tries to merge the two records anyway, maybe a warning could be displayed similar to the warning when someone tries to merge two records marked "not a match." Something like, "Another user has marked these two records as not having enough information to determine whether or not to merge." (Not these words, but this idea.)
I and others I know spend a huge amount of time repairing bad merges, so I would love to see better prevention. W Andrews' suggestion seems like it would help a lot.
1 -
I have a Research Hint to attach a marriage certificate to an ancestor. The hint is correct but I already have the marriage certificate attached to my ancestor. I do not want to add duplicate source as it clutters up the list and makes it more difficult to review my sources for completeness and accuracy.
Right now the option from Research Hints are: 1) Attach or 2) Not a Match.
Please add a third option: Source Already Attached.
Thank you.
0 -
If the system is working correctly, no such option should ever be needed: the hinting system does not normally suggest sources that are already attached. If it does, then either there has been some sort of glitch or error, or it's not actually the same source, just a very similar one (due to multiple filmings and/or indexings of the same register, for example).
If you could give some examples of profiles where you wish for this "already attached" option, we could help you figure out what exactly is going on.
0 -
Here is my example: for Anna Katherine Schopf: LLQ3-DJR
https://familysearch.org/ark:/61903/1:1:QPM1-TRLF
https://familysearch.org/ark:/61903/1:1:QPMT-HHFG
These both link to the same location for Ann's marriage certificate.
Thank you.
Joe Mott
0 -
@josephnephimott1, those are two different filmings (and indexings) of the same document, so they are not actually the same thing. In this case, they happen to be equivalent in terms of readability and usability, but sometimes, one filming is clearer than the other. In any case, both should be attached, because this prevents erroneous attachments to someone else, and also keeps the hinting algorithm running smoothly.
You can see that it's two different filmings by looking at the Document Information on the index detail page, or by looking at the images themselves. One of them is film 102458971 image 551 of 2393, while the other is film 5015786 image 299 of 1066. Notice how the latter has more background material showing and is slightly askew compared to the former.
1 -
I agree with the comment about creating 'clutter' with sources. Often there are many records associated with the same event, especially for marriages.
The sources might be technically different, but since they apply to one event, it would be nice to be able to group them under a single event. That helps reduce 'clutter' in the source view.
It's not unusual to find a marriage record with no image, then the same record with an image, then an index to the record, the event recorded in a statewide index, the event recorded in a county index, etc. It would be nice to attach all of them, but group them to the single marriage event.
Just my thoughts.
1 -
Thanks for the thoughts and comments.
I completely agree that if there is a state record and a county record and a church record all of the same event, I want them all. They are unique sources and strengthen your evidence.
My concern is that when different indexing events have found the exact same source document, it does not add to the strength of my evidence to attach it twice unless there is new information included. If one is only an index and another is an index with Image, of course I'll take the more thorough, complete reference that includes the Image. And if another index project adds no information, I see adding it as cluttering up my list. What if a particular document was indexed 100 times? Is my research more complete if I attach all 100 references to the same document? I feel like my job is to be thorough in sourcing documentary evidence of my ancestors, but my job is not to be thorough in referencing redundant indexing projects.
Just my 2 cents.
1 -
One value that multiple indexes of the same document can have is to resolve handwriting-interpretation questions: if three saw Lemuel and one saw Samuel, it's probably correctly Lemuel.
Now, granted, once you get past a certain number of different indexes (or better yet, indexes of indexes), everything kind of loses its utility, but hopefully they're the outliers. FS staff in charge of creating new indexing projects does pay attention to not repeating effort needlessly. (They haven't necessarily always been successful.)
0 -
Another reason for adding ALL the different, but for same event, is to MINIMIZE duplications in the FSTree. (One of the original 3 goals of FS - Sources, Collaboration, and avoiding Duplication.)
0 -
Yes, I think that if they allowed a 3rd option when checking a research hint- that said yes this is the correct person, but is a duplicate source, then perhaps it wouldn't mess up their algorithm or create duplicate people in FSTree.
Don't know if that's possible, but it would be helpful to have all the correct documentation, without having too many duplicate documents.
0 -
My vote: 3rd category; leaving it to come up again but with a WARNING: Not enough information. Then if they find additional information and add it, like I have, they can merge the two.
2 -
Paul W; Yes to "Are you suggesting to the developers that these hints should remain attached with a clear label, to warn other users (and yourself, for a later time) of further investigation being necessary before making another (or "final") decision on the likelihood of a match?"
0 -
No upvote from me. If you cannot decide, just back out and leave it. Clarity will come later.
3 -
Needs more information to merge
Based on comments by others, I revise my initial suggestion;
When merging duplicates, there should be three options:
1) Merge
2) Not a match
3) Needs More Information.
When selecting option 3, include these features:
1) A text box allowing users to enter comments. This will help others (or yourself later) to know what evaluation has already been done to determine that there is not enough information. Such comments should reside with both persons, and not with the primary persons. Such comments should also reside with each person in the COLLABORATE notes and discussions.
2) An alert whenever new information is added to either person. Or, at least an option to receive an alert. This option should be on an individual basis rather than getting notifications for every person you identified as “Need More Information”
3) Under “Tools”, “Possible Duplicates”; provide a “flag”
“Possible Duplicates (1) Needs more Information”.
“Possible Duplicates (1), meaning that there is one possible duplicate that needs more info and 1 that has not been evaluated.
Having the option of saying “Needs more Information” saves a lot of time. If someone spend an hour or more in research to determine if these two are the same, then goes “Back”, all his/her efforts have been for not. Then the next patron who comes along and tries to merge will spend an hour researching to no avail. And then the next patron, etc. This is valuable time that could be used more productively. Keep in mind that most Family Search users are not retired with nothing better to do. Most family History users have jobs and families and precious little time to do genealogy.
Furthermore, if there are only 2 choices, merger or not a match, there is a greater risk of an error.
If two persons are declared “not a match” then the “Possible Duplicate” notification goes away. If this is done in error, the person that was indeed a duplicate may never be found again.
If two persons are merged in error, the error may never be discovered and/or may be difficult to undo.
(Comment by PaulW: “I and others I know spend a huge amount of time repairing bad merges, so I would love to see better prevention.”)
And, as suggested by PaulW: “Also, if someone tries to merge the two records anyway, maybe a warning could be displayed similar to the warning when someone tries to merge two records marked "not a match." Something like, "Another user has marked these two records as not having enough information to determine whether or not to merge." (Not these words, but this idea.)”
2 -
I agree that there needs to be a way that I can proceed with temple work, even though there is a potential duplicate that I cannot tell if it's a duplicate. Now, I can only mark NOT A MATCH. But maybe it is -- I just can't tell for sure.
0 -
@Julie Melville Hite if you really, really are convinced that these are highly unlikely to be be duplicates, then just mark them as not a match, proceed with temple work, then go into the research helps page to the dismissed hints tab and click might be a match so that the possible duplicate shows up again for someone to continue researching to determine if they are or not. Or just accept current policy and let the work wait until the possible duplicate is correctly resolved.
0 -
If I cannot make a determination I back out of the merge tool and leave notes on the two profiles.
1 -
I agree with many of the above suggestions that there needs to be a third option besides "Duplicate" or "NOT a match". So often I cannot prove that the two parties are indeed duplicates but I am not convinced that they are "NOT a match". Please give us another option that will encourage others to see if they can determine the problem, that someone else could not. I feel I have marked too many people as "NOT a match" for fear of merging the wrong people, while they actually should have been merged, and have caused duplication in temple work. If I simply ignore the issue and just go back and leave it as a "Possible duplicate," then their temple work is denied them.
Could there be a place (Box) where we could leave notes that others would immediately see when they first look at the profile, like at the top? I have often left notes in "brief history" but I always feel bad doing so because these notes are NOT part of the person's history and only my comments or suggestions (maybe even frustrations). How do I let others know that there is an unsolved problem, with possibly other things besides possible duplcates?
Perhaps FS has a better solution that could help in these "tricky" situations. I would love to hear what the experts feel I should do when I am not fully convinced that either option is the best one.
Thank you for your time. I do not know if answers are given to these comments or not, but I am grateful to know that someone will read this and consider my struggles.
Sincerely, Christine Barron
0 -
@BarronChristine1 The good news is that the new person page has a feature that is almost exactly what you are asking for
Could there be a place (Box) where we could leave notes that others would immediately see when they first look at the profile, like at the top? I have often left notes in "brief history" but I always feel bad doing so because these notes are NOT part of the person's history and only my comments or suggestions (maybe even frustrations). How do I let others know that there is an unsolved problem, with possibly other things besides possible duplicates?
The feature is called Alert Notes. It is simply a Note, as has been available for many years, but it has a new option to mark it as an Alert. At the top of the person page for a person with an alert note, you'll see a banner that strongly encourages all users to look at the note details before making edits. This same banner will be shown on the merge screen to make sure people are aware of these notes before doing a merge.
In this alert note you can add all sorts of helpful details about why two people might be duplicates, even if you're not sure. The existence of an alert note doesn't stop anyone from making edits or doing a merge, but it can be very helpful in giving them the background they need to make a good decision before making changes.
Personally, I think the alert notes feature is sufficient to cover the request expressed in the original post in this conversation.
3 -
@W Andrews you said 'If two persons are merged in error, the error may never be discovered and/or may be difficult to undo.'
I actually find a number of duplicates that were marked as Not a Match because I work on one and find a source that is already attached to the other. So, I think these would be found as time passes.
You also said '1) A text box allowing users to enter comments. This will help others (or yourself later) to know what evaluation has already been done to determine that there is not enough information. Such comments should reside with both persons, and not with the primary persons. Such comments should also reside with each person in the COLLABORATE notes and discussions.'
YES! I think there would be tons of benefits if there were more correlation to the Collaborate tab.
2 -
It would be nice if there was a third choice of what to do with record hints. "Undecided" would be a great third choice for those hints that have been reviewed but cannot be determined yet. It would unclutter the Task List and be appreciated.
0