The main reason for this was that we've gotten some feedback from people working on new sequencers, and the SPI being developed on the federation branch had enough benefits to warrant using them in the sequencing framework. For example, one requested change (
http://jira.jboss.com/jira/browse/DNA-77) is to add more methods so StreamSequencer implementations can record properties using paths and the actual data types (without having to convert to Strings). There are also a few bugs with the 'org.jboss.dna.common.jcr.Path' class used by several of the existing sequencers, which has been removed and replaced with an interface and implementation in the new SPI.
Needless to say, there will be in the near future some good improvements coming to the SequencerOutput interface, and the existing sequencers will be changed to take advantage of these.