Conclusions

This paper has described a representation for overlapping hierarchy which is also capable of representing changes to text and attributes. This makes it suitable for some important use cases for overlapping hierarchy, particularly the representation of change between two or more variants of a document.

A significant advantage over some previous representations is that it is pure XML, and therefore can be processed using standard XML tools. The dominance of one hierarchy over another does not need to be fixed and this means that the actual hierarchy of the overlapping structures can be determined for other reasons and indeed varied throughout the document. This flexibility allows fragmentation of elements to be kept to a minimum.

The underlying data model is based on document variants and therefore is better suited to situations where the number of variants is small. Although it does scale to any number of variants, its complexity increases as the number of variants increases, e.g. each new variant has an identifier in the dx attribute so this will become longer and more difficult to interpret.

Overlapping hierarchy is a powerful tool to use in certain markup situations, though its use can lead to complex situations and any solution is also likely to look complicated. This paper is intended to contribute to the discussion as the XML community continues to strive for a simple, generic and universal solution to this problem.

An earlier version of this paper was presented at Balisage 2016 [12].