Skip to content

Category: HyperGlossary

… Further Notes on a ‘Add Term’ dialog box

Further to Wednesday calls it might be useful to let the user choose which pre-set Categories sets to import from the chosen Knowledge Graph. Also, the creation of relationships here is done in a pop-up box where [OHS] [builds] [DKR] is manually ‘coded’, where the verb can be typed or chosen from a list based on the knowledge graph space interacted with.

This is based on the previous post.

1 Comment

Notes on a ‘Add Term’ dialog box for document centric use, which can be shared with knowledge graph systems

A simple wordpress dialog slightly modified to cater for glossary terms, with the option to provide more than one term for search/listing and a free-form text definition entry field which is constantly analysed to provide possible links to previous terms are shown below, for the user to choose to connect to or not:

Analysis For Connecting (?)

I am a very simple guy, with no real knowledge of AI or anything like that but if we really can have the system do basic semantic analysis then it would be very useful I think, so that users won’t have to manually make links between terms (it’s time consuming when you link two new terms and you have to go back to the first to manually create a link to the second and so on). IF the system can extract ‘term’/nodes from text (and go back and do this on older entries for auto-linking and then use the rest of the sentence for link-typing, that would be very useful I think.

Furthermore, if this analysis was in real time with writing the terms then the system could show the old connections found and allow the user to approve, disapprove or modify them. THAT would be great, as mocked up here:

  • Field for Term on top (with ‘+ Add New Term Text’  option to add acronyms etc.)
  • Definition Field in plain text
  • ‘Found Related Terms’ with check boxes to approve
  • Other standard WordPress fields, such as ‘Categories’, ‘Tags’ and ‘media’


The usefulness of this approach of using glossaries will come when someone reads the authors texts and chooses a ViewSpec to access the glossary entries from (stretch text or mouse over fx) and when the author publishes text by choosing to export related terms and ‘freeze’ them in the document for future reference.



The resulting information can then be brought into visualisation systems and presented in a myriad of ways, as a part of a dynamic view in the authoring or reading systems or in full knowledge graphs, such as this:

or even this:

or this:

This continues in the next post.


Glossaries Glorified into Core Connector MacOS Implementation

The way I plan to implement the glossary system for Knowledge Graph integration is in my current systems: 


When writing any document anywhere, select a phrase to turn into a Glossary Entry by launching Liquid and choosing ‘Glossary’ (g) which takes the selected phrase and presents it in a dialogue with options for plain text and single line fields for specific natural language phrases such as ‘xx is a part of yy’. This dialog must be designed in collaboration with the team (Jack, James and David, at a minimum). 


On ‘Enter’ the information is posted as a blog post to a WordPress site with the correct ‘Category’ to assign it as a glossary entry (Liquid already has the ability to post to WordPress). 


The team must then design a scraper to work to take these blog post glossaries and turn them into Knowledge Graph. 


On reading in WordPress, we can have our Hyperwords plugin (which works by showing a blue dot when selecting text) include an option to ‘Show Glossary Entry’ (we already have search and references and so on) which shows the entry in a window. This window will have an option to ‘Open As Knowledge Graph’ which results in a Knowledge Graph software product (DebateGraph, TopicQuests…) opening the associated Knowledge Graph with the current phrase highlighted/in the centre. 


This is only using my stuff, with a minimum of effort to start testing. Other platforms and other software is for others to put effort into but I’m very happy to work on this have have the components and APIs be open, so at least we can try out this kind of a concept.  


What do you think? 

Leave a Comment