An open-source, Python implementation of a Versavant engine, "Versavant.py" is available at http://sourceforge.net/project/showfiles.php?group_id=136572&package_id=150168 and at http://www.versavant.org/releases/index.html. It can be used with predesigned TMAs. Users can also design their own TMAs, using either predesigned TMA component functions, or their own functions, or both. Topic maps that have been created using Versavant.py (either from scratch or by merging other topic maps) can optionally be exported in such a way that they encapsulate their governing TMAs. This makes them functionally self-describing and ready-to-merge with other topic maps; they are examples of "Smart Content". One of the features of Versavant.py is that its operations are easily audited; you can audit the chain of events that led to each property value of each subject proxy.
Versavant-style TMA disclosures allow Versavant engines to provide the service of "semantic integration"— achieving a condition in which there is one subject proxy per subject, with everything known about each subject accessible from that subject's proxy, and in which all of the subjects that each TMA regards as being implicit have been made explicit (i.e. have been represented by their own subject proxies). In Versavant, meta-subjects, including TMA disclosures and their components, receive the same semantic integration service.
A recent paper describes how subject maps can simultaneously reflect "Multiple Subject Map Patterns for Relationships and TMDM Information Items", using Versavant-style TMA disclosures. (backup copy of this paper)
Two posters about Versavant were posted at the XML 2005 conference in Atlanta in November, 2005: