The document has been tagged as being in the field of computer science so any hyperGlossary entries are expected to be related to that particular field of computer science which the author has expressly stated that she is a part of. Joe can choose to use these pre-tagged hyperGlossary entries or use his own specific chosen computer science hyperGlossary. This means that basic (included/pre-tagged) glossary entries can be shown when required but they can also be instantly opened in the readers chosen knowledge graph software and chosen area, for widely useful multidimensional views to really grasp the specific issue, where both the power of visualisation and richly deep AI augments his approach. (I do realise that this leaves a huge amount of issues to be resolved)
Later on Joe starts to work on authoring a new document and adds definitions to the hyperGlossary in a format compatible with the knowledge graph space he is associated with when necessary and explicitly adds previous terms (or denies associations where he expects ambiguity).
When he is done he goes through a ‘Polish & Publish’ stage where he does a grammar check, reader level, plagiarism check, has an automatic outline generated etc. and finally, the software presents terms in his document which could fit with the associated knowledge graphs for him to approve or disapprove.
Joe then publishes the document as a Rich PDF or a Complete HTML document to the repository which is able to extract all the richness in the document and serve the next reader as richly and interactively as possible.