Typesetting multiple interacting streams

We present a new means for specifying multiple interacting streams, as is needed for documents with multiple systems of notes, side-by-side translations, and critical editions. Each stream is treated as a sequence of components, and anchors are used in the concrete syntax to define reference points used by other streams. When these streams are loaded into memory, the anchors simply become iterators in a container. We present a set of algorithms for the typesetting of multiple streams of text, each with multiple streams of floats and footnotes.