[desmap]

So, after diving a bit deeper into this field I still do not understand why the underlying tech is important for bidirectional linking?

I understand that a graph database like neo4j can be amazing once you need more than just binary relationships (any kind of relationship in any level can be expressed) and once you start to query the database. Latter is so powerful and crucial from a data scientist perspective.

But why are these graph db features important for notes? And what are the drawbacks if you just do this with normal text or md files with links and a normal file system? Assuming that the actual link setting UI is convenient and takes care of most there shouldn't ne any disadvantage when writing and querying them. If things get slow you could still use an index but yeah.

One advantage with normal files is that you will still be able to read them in 50 years without any special app created 50 years ago.

[tincholio]

You can have both, actually. Having a datalog engine allows things like this, which is pretty cool: https://www.loom.com/share/d007932e94db4b4981cca606bebdb54a

[desmap]

Thanks for the video. Still I am not sure why you need a graph database here, everything can be done with hyper-linked md-files, optionally enriched with metadata (front matter). I am aware that a graph database is more elegant but again, does it justify its costs?