Semantic knowledge

#1

I want to share Hode, a Reflexive Set of Labeled Tuples or so called hypergraph. What is interesting is the syntax with three elements:

a phrase (like “cats”), or a relationship (like “cats have noses”) or a template (like “_ have _”) shared by many relationships

The goal is to organize knowledge more like natural language; e.g., “birds #eat worms” can be used to query “what do birds eat?”. In Hode, one establish such relationships with “(Saudi Arabia #exports oil) #(because, I’m guessing) its extraction cost is 3 US dollars per barrel”

The idea seems related to ORM (Object-Relational Mapping) as investigated and tested in TiddlyWiki by Charlie:

This is also related to node mapping/graphing, such as by TiddlyMap.

TiddlyWiki can already establish relationships between tiddlers, by tagging, linking and fields. Even more would become possible if [semantic annotation of links] (In-text semantic annotation of links · Issue #2144 · Jermolene/TiddlyWiki5 · GitHub) is implemented. Or alternatively, inline field definitions (in the tiddler bird, writing “Many birds eat::orms” populates the field eat with orms)

Just some food for thought!

Also related: Semantic Synchrony

2 Likes
#2

In a similar manner I have started investigating “sentences” found in text followed by “?” as a question and similar relationships (?! !?..), but then also parse text for verbs (doing words) to identify actions or “todos” from text. As recently shared we can use transclusions to capture a relationship from here to there.

Given this thread I will look to see if there is a similar method for “something relationship something” in effect a 3 parameter relationship definition, or Tuple. Once we have this we may be able to institute a way to query these.

I will follow up the other resources you mention, thanks for raising good “strategic questions”.

#3

I think this is a good idea because it reduce the effort needed to input thing into fields to achieve: Suggestion for all Plugin developers about field name: Let's use Ontology to maintain Interoperability

Natural language + markup make it easy to input tuples. And to make the plugin developers easier to use these tuples, we still need them to be in the fields. (imagine you are a plugin developer, how will you use data?)
Maybe there can be a plugin to auto-extract these sentences into JSON-LD and put them into fields.