The above-linked post and other discussions on this forum hint at the fact that we all use nesting (of and in tags and groups) to mean several different things. Got me to wonder why.
My main DB shows a very dense usage of group tags that builds a dense and interconnected/crosslinked topical/concept map, onto which I connect (by mean of tagging, classifying/replicating in a group) different documents.
In building such a map I inevitably have to make tradeoffs: how do I represent (through a nested hierrarchy with repllications) that a topic is at the same time:
- a precision of another one
- a component for another topic
- the opposite of a topic
Topical hierarchies have their limits. Limits that Ontologies have not.
And here comes my suggestion : a killer feature for users with the need of topical density would be to add the support of a third mean of “categorizing” document : ontologies.
If we look at how it could be implemented:
- ability to define relationships between tags or groups (aka build a knowledge graph).
- ability to define the set of possible relationships (aka an ontology) for a DB, with a good default one being SKOS.
- ability to visualize and navigate this graph of topics.
I know such ontologies and knowledge graphs are more and more in use in diverse sicentifical fields so I consider it as having a great potential for new users: create a DB, load the right ontology and bingo! you got a great classification structure.
“Aside tags and groups”. I know such power-tools are not useful for everyone. It can get easy to get lost in such things. That’s why I think it should be a “third way”, beside groups and tags.
Just a thought that I needed to deposit here.