Photos are a good example of why a filesystem hierarchy is insufficiently expressive. You might want to search for pictures of me at parties, or pictures with me and my wife, or pictures from 1999, or pictures of LA, and the same photo might belong in all of those searches. No single category will ever be a good place for a photo.
That's a good example because it leads to truth that NO up-front tagging will ever anticipate all the searches you might make in future. There are so many possible searches.
I figure that a combination of wetware and software is the current sweet spot. My brain usually has enough associations and context to turn every photo search into a time or place filter - "I think it was downtown last year" or "some time in summer at home" or "it had my wife in it". The photo storage system need only provide search/filter on date and place to narrow it down to a few hundred thumbnails, plus machine-learning to tag people. Which is basically what iOS provides, no more, no less.
Any other up-front categorization or tagging is basically wasted effort.
For me, I wrote a bulk tool that renames my photo file names by reverse geocoding the GPS information via Open Street Maps. That way I can do text search for place, as well as 2d map search. It's at https://unto.me
I’ve been looking for a good tagging system for image and video files, something I can use to quickly and easily go through a stack of files and tag them, then search by tag later. Bonus points for being able to recompress on the way through, since phones seem to have terrible compression ratios compared to offline compressors.
That does seem like a really good solution. Google Photos will prompt you, if I recall correctly, to identify a few faces, and then automatically id the rest. That's fantastic, if you don't have to worry about putting their privacy in Google's hands.
I personally agree with your point (and find the loose textual search offered by phones these days to be mostly adequate).
But reading your comment gave me a thought: filesystem hierarchies are indeed insufficient, but what about filesystem hierarchies with liberal use of hardlinks?
That seems equivalent to a graph to me, and yes, I'm unaware of any kind of search that a graph does not permit. Indeed it could be the basis for a system that, in my opinion, would dominate any of the existing knowledge graph / tool for thought products. It would consist of three more pieces:
* A database for backlinks. (Links from file X to file Y would only be possible when X has an appropriate file format -- `.txt`, `.md`, `.org`, etc.)
* A search grammar with the following primitives:
* find children of (links from) query results
* find parents of (links into) query results
* take the disjunction (OR) of queries
* take the conjunction (AND) of queries
* group queries with parentheses
* The ability to pipe files found via ordinary shell commands into that grammar.
Given the size of most peoples' knowledge graphs, you wouldn't even need to keep a text index (ala Lucene) -- `find` and `grep` would be more than sufficient.
