DTMNodeList gives us an implementation of the DOM's NodeList interface wrapped around a DTM Iterator. The author considers this something of an abominations, since NodeList was not intended to be a general purpose "list of nodes" API and is generally considered by the DOM WG to have be a mistake... but I'm told that some of the XPath/XSLT folks say they must have this solution. Please note that this is not necessarily equivlaent to a DOM NodeList operating over the same document. In particular:

• If there are several Text nodes in logical succession (ie, across CDATASection and EntityReference boundaries), we will return only the first; the caller is responsible for stepping through them. (%REVIEW% Provide a convenience routine here to assist, pending proposed DOM Level 3 getAdjacentText() operation?)
• Since the whole XPath/XSLT architecture assumes that the source document is not altered while we're working with it, we do not promise to implement the DOM NodeList's "live view" response to document mutation.

State: In progress!!