Directory systems on machines are implementations of set theory. A folder is a set. But it is limited. There are two main questions:
Which mechanisms have been invented to circumvent some limitations currently and is it posssible to expand on these using theoretical knowledge
Which mechanisms have been created to work collaboratively on individual ideas on structure of sets and metadata and is it possible to explaind
on these using theoretical knowledge
For the first one we see e.g. metadata inside files e.g. EXIF or metadata stored in seperate files or external repositories like local datastructures.
For folder we see e.g. on windows the folder.ini and on MAC the DS Store to capture metadata on set level.
For the second one we see e.g. documentation management systems which implements taxonomies, folksonomies using search and many other mechanisms to derive
e.g. legal properties or security properties from the structures.
For relations we see shortcuts to internal resources and url files to link to external resources. Which is currently not clear if it is ment "see also"
or "points to canonical taxonomy like UDF".
Interesting would be if hooks could be created to advise the user on structure and metadata while building a new directory structure
I think the idea of building a local directory structure or a company having millions of documents on a shared drive and coming up with only the
solution of a document managment system which abstracts on top of that in another system is not logical. I think we can better extend the existing
file managers to be part of a global structure. So that every individual folder structure on every machine is always part of also a global
folder structure and the files themselves are handled by something far more web based.