The current version of the NXT documentation is in directly-styled XML format. If you use Internet Explorer, or a very old version of another browser, please use this HTML version which has the same content but weaker navigation, or download a new browser (this site has been tested on Firefox, Opera, Chrome, Safari).
You can also download this documentation as a single pdf file.
In early 2011 we ran a couple of tutorials on NXT. Tutorial 1 covers some NXT basics and introduces the query language - it's intended to give you an introduction to the skills you need to investigate an existing NXT corpus. Tutorial 2 introduces various skills useful to those creating or managing their own corpus in NXT format, or creating new annotations on an existing corpus. Though both tutorials contain some paths and commands that are local to us in Edinburgh, we have added comments in italics for remote users so you can achieve the same effect.
NXT documentation is authored in the DocBook XML dialect. This allows web and print versions to be created from the same source with minimal effort. XML can now be styled directly with CSS in most web browsers, and this avoids complex transforms. For web presentation we apply a simple XSLT transform to split up the DocBook into multiple files and add navigation, then style with CSS based on Michael Thiele's wysiwyg DocBook stylesheet. For browsers that do not implement these standards, the transform to HTML is done using xmlto. For PDF / print output, the tool of choice is Prince. There are a few remaining issues with links in the rendered PDF version but it's better in many ways and provides a more flexible approach than the xsltproc / xmlto approach we previously used which goes via XML flow-objects.
All documentation XML files, transforms and CSS files used to create the documentation for print and web are available in the Sourceforge CVS tree in the doc directory.