Wednesday, September 22, 2010

Relatively Eye Popping

If you have any eyes left to see, today I’ll give you a peek at Relationship Calculator. Relationship Calculator is the third of three eye popping technologies  demonstrated by Tim Cross at the recent BYU Family History Conference. Cross is a product manager at FamilySearch, working to develop a genealogical community leveraging the new FamilySearch (NFS) Tree.

Relationship Calculator

In several places around Temple Square one can see descendancy charts showing common origins of famous people. One chart shows descendants from the Howland family. Another shows descendants from the Loomis/White families.

Descendancy tree in the FamilySearch Center of the Joseph Smith Memorial Building
Descendancy chart of Robert White and Bridget Allgar

I thought one of these relationships would make a good test case for Relationship Calculator. The charts are of uncertain quality, to be sure. And the technology is still under development. So I didn’t expect everything to work flawlessly.

On their Tree Seek website, Misbach Enterprises provides a simple interface to Relationship Calculator. They call it “How are we related?” I used it for the test. (Sorry, friends; it requires an NFS account.)

I thought it best to pick two people that were famous—but not too famous. That would minimize the possibility of IOUSes blowing up the calculation. I first tried one of the most distant relationships. Orville Wright and Philo T. Farnsworth are 8th cousins, twice removed.

You know Orville Wright. Philo T. Farnsworth is the inventor of the television and is honored by Utah with one of its two statues in the U.S. Capitol’s Statuary Hall Collection. (The other is Brigham Young.)

I tried LW1S-2J2 (Wright) and KWCQ-PFY (Farnsworth) in Relationship Calculator. It could not find a relationship.

Could unresolved duplicates in the NFS Tree sever the connection between the two?

I checked Farnsworth and was able to successfully navigate back to Robert White (KNZM-D6J), as shown in the chart below.

Philo T Farnsworth is a descendant of Robert White and Bridget Allgar

But I found two points where the default parents were incorrect. Perhaps that caused the failure. I checked to see if Relationship Calculator could connect Farnsworth with White, his 9th great grandfather.

Relationship Calculator decided they were 5th cousins, 3 times removed. Oops.

RelationshipCalculator

I’ve copied the names into the table below and added birth and death dates.

Common Ancestor [not specified]
Siblings John Cary (1583-1661) Robert Carey (1457-1540)
1st Cousins Sarah Morgan (1700-1777) William Cary (1482-1528)
2nd Cousins [not specified] First Lord Richard Cecil (1495-1553)
3rd Cousins Ruth Harmon (1733-1755) [not specified]
4th Cousins Lydia Shelden (1759-1846) Richard White (1516-1578)
5th Cousins Lucinda Kent (1785-1829) Robert White , Jr. (1558-1617)
  Philo Taylor Farnsworth (1826-1887) —once removed—
  Lewis Edwin Farnsworth (1865-1924) —twice removed—
  Philo T. Farnsworth (1906-1971) —3 times removed—

 

Oh boy.

A little pedigree analysis shows that John Cary (1538-1661) and Robert Carey (1457-1540) can not be siblings. Nor can Sarah Morgan (b. 1700) be the daughter of John Cary (d. 1661), nor can Lydia Shelden (b. 1759) be the daughter of Ruth Harmon (d. 1755), nor is it likely that Richard Cecil (b. 1495) is the son of William Cary (b. 1482), nor is it likely that Richard White (b. 1516) is the grandson of Richard Cecil (b. 1495).

Is the problem Relationship Calculator or is it the data in the Tree? As the old saying goes, “Garbage in, garbage out.” (GIGO)

Whether Relationship Calculator works or not is immaterial while the tree is full of garbage.

 

Next week I compare Ancestry.com’s Famous Relatives Finder. Can it do any better? Stay tuned…

No comments:

Post a Comment