Thoughts on Resolving the Family Tree Dilemma

In my last post, I reviewed some of most perplexing issues faced by the FamilySearch.org Family Tree program. See "The Persistent Merging Problem on the FamilySearch Family Tree." In this post, I will conceptually examine some of the solutions to the problems.

The FamilySearch.org Family Tree is a wiki designed to hold genealogical data. As such, the Family Tree is structured so that each person who ever lived on the earth can theoretically occupy only one node in the tree structure. Since each person who has ever lived can only have two biological parents, the tree structure is a lattice of interrelated individuals. A single node is the product of two individual parents, however, a node can also be "related" to additional sets of parents based on social conventions, such as adoption, foster child, guardianships etc. If there were no such alternative relationships, the tree would be 2 dimensional. In order to accommodate these alternative relationships, the tree must become 3 dimensional and allow links from an individual to more than one set of parents. In addition, every individual represented by a node can further be related to other individuals who share the same set of parents. The tree must also represent culturally defined marriage relationships between two or more individuals. An individual may have multiple marriage and marriage-like connections. Even though the tree provides a node for two or more "parent" relationships, some of these individuals may not be identified even though, theoretically, there is a place in the tree for all of these unidentified individuals.

For someone designing such a program, there is a fundamental challenge of adequately representing these possibly complex relationships in a way that can be visualized. The presently designed Family Tree does an adequate job of representing the full, potential lattice of relationships. As a user, I can start with myself and navigate to anyone who is related to me in any way, including all of the non-biological relationships. From the standpoint of processing all this information, computer systems that can adequately store the entire lattice are just now beginning to be available. The adaptation of the Family Tree to a computerized representation is made even more challenging by the need to document every single node (individual) on the tree structure. Because the number of related ancestral nodes increases geometrically, the number of potential nodes is possibly over 100 billion individuals.

The tree structure also has to allow for the reality of pedigree collapse when individuals marry relatives with a partial copy of their own ancestral lattice.

Given these parameters, those who have designed and developed the Family Tree have done an amazingly effective job. Now, I need to consider the data or information incorporated into the tree structure. As I have pointed out in many previous posts, the challenge of incorporating the existing data into the tree dwarfs the challenge of creating the structure in the first place, especially when you consider the fact that the information is duplicative, incomplete, contradictory and some is completely fabricated. In addition, the "users" who are being allowed to contribute even more information to the tree range from extremely sophisticated to entirely clueless. Here the collaborative structure of the wiki excels. The wiki structure allows an infinite number of "corrections" to the initial data entered into the system. In the case of the Family Tree, the combined data presents a variety of selection options. Since only one individual can ultimately reside in his or her respective node, all of the duplicates must be either deleted or ignored by the program. By its nature, the tree will incorporate incorrect information and the node structure will not accurately model reality. But because of the ability of anyone to make changes, the tree structure will inevitably evolve into a more accurate representation of the real relationships between the nodes as the information is corrected and the relationships more correctly defined.

To summarize; the structure of the Family Tree is an adequate representation of historical reality. It has the ability to incorporate the full spectrum of family relationships and adsorb all of the information attached to every individual. As the program evolves, it will inevitably begin to incorporate any additional structural needs that can be identified.

Users of the program come with differing amounts of initially possessed data and opinions. Some users come with little or no knowledge of their family and its relationships. Others possess huge files of information. The Family Tree accommodates both ends of the user experience. The open, collaborative structure of the tree is foreign territory to most users. Genealogists are usually solitary researchers with a fixed, preconceived view of the data they have collected. Getting these individuals to cooperate and accommodate those who they may disagree with is a major challenge. Those users faced with accommodating conflicting entries often become defensive and have the attitude that unless the game can be played their way, they will simply refuse to play at all. Fortunately, the tree structure of the program simply ploughs around these rocks in the field and moves on.

What about the huge pile of duplicates that were entered into the program by seeding it with multiple databases accumulated over more than a hundred years? The mechanism for resolving all of this information is built into the program. Practical limitations on the access to certain parts of the information, i.e. Temple related and membership related information, makes the resolution of the duplicates more complicated than simply opening up the program to a massive merging effort.

What is the most effective methodology for working with the Family Tree? The basis of any progress in sorting out the information in the Family Tree is careful, documented research. As I have added sources and corrected the existing information from those sources, the Family Tree has become locally more reliable. This is a slow, methodical process. Not every data field can be completely established and/or verified. By their nature, historical records are often conflicting and may be inaccurate or incomplete. Some relationships may never be adequately established. Ultimately, the only way the Family Tree overcomes the merging problem and every other challenge to the integrity of the information is step by step and data field by data field.

Every user of the tree needs to closely examine all of the sources added to every individual for relevance. Inappropriate sources need to be detached. The information presented by the sources needs to be evaluated and where appropriate, used to correct the entries in the Family Tree. All unsupported relationships need to be deleted. As the program progresses, all duplicate entries need to be merged. This needs to be done, individual by individual, systematically working backward from the individual user. Any conflicting information needs to be resolved by collaboration. As we collectively work on the Family Tree we will soon find the current limitations of the information and can begin to add more information that is missing. As a byproduct of this methodology, new individuals will be added to the Family Tree. Skipping generations back into the past to do "research" on remote ancestors is counterproductive. Unless all of the intervening individuals are adequately documented, the users who jump generations are almost guaranteed to be working on people to whom they are not related.

We are not so much building a new structure as remodeling an old one. We need to make sure of the adequacy of the construction of the tree at every level before adding on. Patience, accuracy, responsibility and most importantly love for the people in the tree should be the basis for what we do. To quote from President Gordon B. Hinckley from the August 1986 Ensign,

Rather, let us go forward with faith and with the vision of the great and marvelous future that lies ahead as this work grows in strength and gains in momentum. Build faith in the hearts of all those around you.

Quoting further from the same article,

May the Lord bless us as builders of faith. May our testimonies strengthen and become as anchors to which others may secure their faith in hours of doubt and concern. May the candle of learning ever burn in our minds. Above all, may testimony grow in our hearts that this is in reality the church of the living God and that it will continue to move forward to fulfill its divine destiny. May we each do our part faithfully and with thanksgiving to the Lord for all the blessings he so wondrously bestows upon us as we follow his teachings and draw near to him.

Let's get to work.