Alignment qualifiers

This limitation may be addressed by alignment qualifiers – metadata modifying the interpretation of the alignments. In particular, the propagation, distribution and atomization of the source context might be adapted to special requirements not covered by the default rules. This approach requires a metadata model defining the names and semantics of “aqua items”, metadata items playing the role of alignment qualifiers. Embarking on this task, one should remember a basic difference between the SNAT model and the SAAT model: the SNAT model is based on expressions and closer to a software developer’s perspective; the SAAT model focuses on alignments which can be identified by a subject matter expert. Simplicity is an important benefit of the SAAT approach, without which it might not be worthwhile, considering the superior expressiveness of the SNAT model. The definition of a SAAT-based metadata model may be perceived as the challenge to find a tradeoff between simplicity and expressiveness.

An important usecase of SAAT models is the support of graphical mapping tools: the graphical representation of a document to document transformation can be captured by a SAAT tree, which can in turn be mapped to a SNAT tree providing both – a specification of the mapping semantics and an implementation of mapping source code.

The construction of a SAAT based metadata model is work in progress, geared towards supporting a graphical mapping tool which is part of a framework supporting data transformations between large type systems.