We're happy to discuss future development work with users and potential contributors. We will also consider undertaking some developments for pay but are much more interested in this when the results remain open source.
There is an active NXT SourceForge page where bug report and feature request lists are maintained. Some of those that are planned for implementation are listed here.
The layering model imposed by NXT's metadata on a data graph sometimes gets in the way of what users want to do. It would be fairly simple to allow pointers to draw targets from a disjunction of layers, or to allow individual layers to be skipped, overcoming the main difficulties they have reported.
The layering model imposed by NXT's metadata on a data graph is one simple way of avoiding cycles in the parent-child relationships that users can easily understand, but it is more stringent than the NITE Object Model requires. We might also consider removing the layering model completely. We would like to see experiments done with an alternative metadata format, data storage, and loader that do not use layers. This would take careful design if we are to retain the advantage of multi-file serialization.
Although it will never be as useful as tailored tools, the Generic Display could easily be improved. The Javadoc comments at the beginning of the code for it give our ideas about that, and it would make a nice student project.
It would be useful if NGramCalc could allow order to be derived from end time rather than start time, or from a structural order rather than a temporal one.
Last modified 10/17/07