By exportation of a document, we mean the deep copy of all the document content to an external repository. Exportation and distribution of paper Customer Documents are equivalent: apart of the external reference list, usually provided as an indication to the document source, paper documents contain everything the reader need. And the only way to distribute a paper document is to copy its whole content.
HyperText documents are very different in the way they use references (or links), which are provided in HTML files as garantees to an existing and up to date content. Clicking on a link will bring that content to the screen. Exporting a document should provide the same facility.
Unfortunately, because of the required absence of implementation structure of HTML document bases, exportation of HTML documents, or even of document parts cannot be generic, and the worse should be assumed: exportation of any Web page may potentially lead to copying the whole Web, including the whole Internet and that is hardly what we want!
Instead, exportation should be limited to only some limited set of links and left to be defined by every document owner or creator in the following way: a distance, expressing the sub-directory level from the base HTML document to any target file could be defined as a recursion limit. Links following could be as well limited to relative links, providing a way not to inline documents external to the server. GNU wget implements these features and could be used as a base for a future implementation. University of Arizona's webglimpse could provide help as well.