As its site says, libfolks is “a library that aggregates people from multiple sources”. Because the modern user has so many different sources of contacts – address book, IM, social networking, etc – duplicates from the different backends are inevitable. We call these (potential) duplicates from each backend Personas, and the goal of libfolks is to link the Personas into non-duplicated Individual objects. Each Individual contains all of the fields from the different Personas that are linked together – the phone number from the address book, a current alias from a Social Network, etc.

Since there is no need to explain the obvious benefits for Desktops in having a library like libfolks around I'd like to dive directly into Folks' architecture and writing new backends. The current version of Folks supports two backends: one is a key-file and the other is based on Telepathy. There is also a third backend for libsocialweb that was developed by Marco Barisione which allows you to fetch your contacts from Facebook and Twitter (and eventually many other social services). Recently, I was asked to implement a new backend to interact with EDS (Evolution Data Server).

The goal of this talk is to illustrate the core elements that compose a backend in Folks. With what we'll cover that attendants should be able to leave with the practical knowledge needed to write their own backends. The second aim is to open a debate in the area of contact aggregation.