Search a person just in your section of the family tree
Ancestry.com has a tab on the page when you are viewing your tree that allows you to search for a specific person just on your tree. It would be great if you could do the same on FamilySearch instead of having to search the whole tree and put in all of their information, like DOB, spouse, and children to narrow down the search results. Would you be able to put a "Search Your Tree" box in the upper right hand corner when you are on any of your tree view options? Thank You
Comments
-
How do you define "your tree" for FamilySearch? There's just the one.
1 -
@Cathy DeCook You have the option on the FamilySearch tree to "follow" up to 4000 profiles. If you follow your immediate and extended family, then you can search within those profiles you follow - even on just part of a name. https://www.familysearch.org/tree/following/
You also have the option to add labels to profiles you are following, to narrow groups even more.
3 -
@Cathy DeCook This has been requested frequently. The programmers here do seem to be quite a creative bunch so maybe one day this ability will be added. The basic problem, however, is computing resources. I'm not all that creative, so all I can think of is that an ability to see search results that just contain current relatives in Family Tree would require the program to first run the FIND routine then the View My Relationship against each name and hide those with no relationship.
For example, if I tried this with John Smith FIND would collect 6,552,925 results so the program now needs to run View My Relationship that many times. Certainly doable, but probably at an unacceptable cost in terms of time it takes to display the results and what it would do to overall system performance when 50,000 people are all trying to do this at the same time.
But as I said, maybe a developer will come up with a much more efficient way to do this that makes it possible some day. They have done some amazing things here.
0 -
For example, if I tried this with John Smith FIND would collect 6,552,925 results so the program now needs to run View My Relationship that many times.
No experienced programmer would write it that way. A much more efficient algorithm would optionally first isolate "your tree", then a hash of the personal names it contains. Dates would be left for last because they are so often not standardized.
But what is "your tree" when anyone can engage with any part of this one tree of all humanity? What this request is really asking for is a tool to search or filter by name exclusively the ancestry tree and/or descendancy tree of any given PID. The partner website Puzzilla has that tool, albeit only for a limited number of generations.
Descendancy trees can be enormous. I am working now with the descendancy tree of one early American colonist. Although still far from complete, it includes hundreds of thousands of PIDs.
1 -
The though had occurred to me that the other option would be to start, as you say, with isolating all my relatives in Family Tree. One way to do that would be to run View My Relationship against all billion+ people in Family Tree and find everyone related to me then run FIND. The other would be to start with me and follow every line back as far as it goes (or to some limit such as the 15 generations the View My Relationship is set at) then follow each of those end of line people forward as far as can be done, then run FIND against those millions of people.
But then I haven't done any real programming since fiddling around with FORTRAN and BASIC years ago so I'm sure there are far more subtle and efficient ways of doing this. But I'm afraid any of them at this point are just not practical. I've played around with Puzzilla and been able to request a chart that takes several hours to generate and others that have just crashed while trying to generate them.
1 -
Traversing the tree is the fast way.
However, as I said, descendancy trees can be enormous. I'm not sure it is something FS would want to implement for the convenience of users. On Ancestry most trees are very small, so searching a tree for a name is feasible.
The other part of this is the frequent request for a more refined date search tool.
1 -
Ancestsry.com has PRIVATE trees - FamilySearch doesnt - THATS the difference - - the average person can have MILLIONS of connection on their tree in FamilyTree - even if they didnt input any of it. FamilySearch FamilyTree is a SINGLE collaborative family tree database of billions of records. versus Ancestry where people each have their own tee.
0 -
Thank you for your comment. I am well aware that FamilyTree is one huge worldwide tree. However, when you click on Tree under Family Tree the program shows just YOUR small section of the Family Tree. Therefore, I thought it might be possible for the program to search just the names you are related to. It can show whether you are related to any person in the Family Tree when you click on Show My Relationship, so it just seemed like the program would also be able to find a specific person that you were related to. Obviously, I am not a programmer, but I just thought it might be a possibility.
0 -
Thank you for your suggestions to use Puzzilla. I was not aware of that feature.
0 -
@Cathy DeCook, yes, it is definitely possible to program such a feature and such a feature has been requested many times over the past ten years. My comments, as an extremely out of date dabbler in programing, was to point out that possible does not always mean practical.
If the designers do start work on what would be greatly welcomed by a lot of users who think they want it, their first decision would be where to set the limit between useless (search just direct ancestors back four generations - you can see all these on the fan chart) and overwhelming (all direct ancestors, all descendants of all direct ancestors, all descendants of all direct ancestors and those descendants' spouses, those people plus all the direct ancestors of the spouses, all the preceding plus all the descendants of the spouses ancestors, etc. - pretty quickly, you have the entire database anyway.)
Maybe what we need to request is a very basic beginning: a search that looks at just our direct ancestor back 30 generations. That might encourage the programmers to start investing the time to look at creative ways to generate such a list. Then they could release it and see how much people actually use it and if they find it beneficial. If it worked, then they might be able to find ways to start adding just the descendants of people on that list.
0 -
just our direct ancestor back 30 generations [...] just the descendants of [that ancestor]
Back 30 generations. Um...
One person in one of my surname studies has about 10 generations of descendants, totaling over 400,000 persons in Family Tree. And the surname study is far from complete.
Every one of those 400,000 persons has 2 parents, who each have 2 parents, etc. Going back 10 generations, every person alive today has 2^10 +2^9 + etc. = over 2000 direct ancestors. Every one of those ancestors has a descendancy tree.
So, traversing the tree as an initial step in a search is not compute-efficient, and also unnecessary because already there is a much more efficient search provided.
The more efficient search provision is this: Simply search by name, approximate birth date or place or any other information you have about the person. Inspect the profiles of the most likely candidates to see how they are related to you. Understand that if the tree is fragmented they may appear to be not related.
Defragmenting trees is the primary goal of a surname study.
1 -
That's my point. If a routine can be developed that effectively searches just for our relatives, we'll still probably end up with such a huge fraction of the database, we might just as well search the entire database.
3
