Streaming ranked-tree-to-string transducers

Streaming tree transducers with single-use restriction (STT_surs) were introduced by Alur and D'Antoni as an analyzable, executable, and expressive model for transforming unranked ordered trees in a single pass. The equivalence problem of STT_surs is decidable because their class is as expressive as the class of MSO-definable tree transformations. In this paper, we present streaming ranked-tree-to-string transducers (SRTSTs), based on STT_surs: SRTSTs are released from the single-use restriction while their input and output are restricted to ranked trees and strings, respectively. We show that the expressiveness of SRTSTs coincides with that of deterministic top-down tree transducers with regular look-ahead (yDT^Rs), whose equivalence problem is known to be decidable. Our proof is done by constructing equivalent transducers in both directions.