Wednesday, March 12, 2008

FamilySearch Developer Conference Notes

I'll be adding some notes throughout the day...

Ransom Love: “Brave New Platform: Changing the World of Genealogy” (It's nice to see that FamilySearch is using plone)

Duane Kuehne: API Overview
They do not yet have load metrics on any of the API calls they outlined. They expect to have an SLA (service level agreement) with this information at some point. One thing that surprised the API dev team was the large size of individual records once you combine all the duplicates into a central location. Some individuals can be on the order of a hundred MB or more.

Ryan Heaton: Family Tree Read

Resource Types: XML, JSON, can be gzipped
Resource Locations: Via ID or Identifying parameters
You can read a person/place by ID, name, etc, etc, etc.

Data Definition
XML Schema Location:

persons - contains data about a person
searches - contains data about searches
users - contains data about users
matches - contains data about matches
personas - contains data about a persona

person (nested in familytree/persons)
  • information (ids, gender, etc)
  • assertions (the data that makes up the person (names, events, facts,relationships, etc)
  • summary (most relevant name, gender, birth, death, spouse, and parents)
  • values (grouping assertions by value)
  • composition (a view of what persona make up this person)
search (nested in familytree/search)
  • this is different from the person data above
  • score (relative to other search elements)
  • ref (the id of the person)
  • person/parent/spouse
  • places
  • dates
  • names
Testing it out (you need to have an API key first):

Get all data on a particular person using their id:

Query for values and summary:

Get user data on a particular person:

Get summary view of 2 generations of ancestors for a particular person

Get JSON data for a particular person (not sure if I got this one right):

Matt Misbach
(Bungee Labs): Rapid web client development/deployment
  • The buider IDE itself is a bungee application
  • No data management included, but this can plug into S3 (it runs on EC3).
  • The IDE will always be free to developers
  • Business model: it will eventually be billed based on a utility model (a combination of server memory footprint, bandwidth, and CPU). For this year (the beta period) it will be free.
  • Create an account and email your username to matt[at]bungeelabs[dot]com, then he'll invite you to his developer group with the bungee FamilySearch API.
  • is an example of what you can do with the FamilySearch API
  • WideLens is a more general example app that combines into a single calendar, SalesForce, GoogleCalendar, etc, etc.

Rob Lyon: FamilyTree Combine/Separate
Tree cleaning (removing duplicates) - someone needs to develop an app to handle this better.

No comments: