So speaking of organizing my Dropbox folder... I'm facing again the whole taxonomy vs folksonomy thing. Taxonomies are seductive hierarchical classification systems that lure you with their siren song of logic, order, and sense-making.
The Yahoo! directory (composed of hierarchical categories) was an early famous example of an attempt at ordering the world's knowledge (well, really just the world's Internet at that time) in a taxonomy that provided intuitive navigability, discoverability, rapid look-up.
Trouble was, it wasn't super scalable, and lots of things refused to be neatly pigeon-holed in a single place with the hierarchy. (Right now I can't remember if the same content could appear in multiple places). File systems (generally hierarchical) face the same problem.
This has been the case ever since you could put a folder or directory inside another, and it exists today in the age of Dropbox. You can, of course, have something "live" in more than one place at a time if you create a symbolic link (or a hard link, or an alias, or a shortcut).
I found that when I used to do that on HFS filesystems, the aliases weren't super reliable and could easily rot over time. Layer a cloud-based synchronizing abstraction like Dropbox and the probability of breakage gets much worse.
There have been lots of attempts at overcoming the pain points of working with large taxonomies by approaching the problems from a different angle. Google aims not to make the world's knowledge conform to a hierarchy, but rather to make it searchable.
Microsoft made a "database" filesystem. BeOS kind of did too, with its rich metadata. Apple did as well with Spotlight (long before APFS). All of these attempt to present you with an abstract view of the volume that elides the structural location on disk.
Gmail is another great example. You can make "folders" and even nest them inside each other, but they are called labels, and you can apply multiple labels to a single message, causing it to show up in multiple places.
The idea is that the search should be so good — and it generally is — that you can find what you are looking for based on what it is and not where it is. This idea of tagging first came up on sites like Digg (I think?), AFAIK one of the early great examples of a "folksonomy".
These tag-based classification systems work great IMO. Unlike hierarchies, which you have to pretty much get right from the get-go (because they are hard to change, because major conceptual flaws are difficult to correct, because some stuff *is* intrinsically ambiguous) ...
...tag-based schemes are easily modified and lend themselves well to incremental improvement. Major conceptual deadlocks are rare. You can *invest* in a tag-based system and it will become more valuable over time. The investment still makes sense even if the search technology...
...is weak, because the latent value of the tagging can be capitalized upon "for free" whenever the technology improves. They're great. Alas, none of this helps much with the question of how to organize your Dropbox, because you still have to decide *where* to put your stuff.
Some illustrative examples: should I create folders for my kids and nest their stuff inside it? (eg. a folder for health insurance, a folder for doctor visits) or should I create top level folders for insurance and per-person folders within those?
What if I have a major medical episode? Does that warrant its own top-level folder, and what should go inside it?
How much of this should I redundantly encode in the name (eg. "John annual medical checkup 2018-09-10.pdf") and how much in the path (eg. "Medical/John/Annual checkup 2018-09-10.pdf")? Which is going to be more searchable?
Also, do you notice how I don't really trust the OS filesystem to preserve the dates or make them usefully exposed when searching or browsing? When should the date go in the filename (if ever), and when should you encode it in folder structure (eg. "2018/10/...")?
Unlike many other areas in my life where I have relentlessly optimized and refined my practices and processes to a point where I am relatively happy with them (no in the sense that they are perfect, but in the sense that they have solid rationales and are reasonably effective)...
...when it comes to file organization, all I have is a massive corpus of artifacts manifesting a many-year style drift where you can see how my patterns and practices mutated over time, never really settling. At this point, the only thing I haven't fully tried is throwing it...
...all in one massive directory and periodically trashing or archiving the oldest files into subdirectories, relying on search and times winged chariot to sort out all the rest. I was kind of hoping that I'd have an epiphany while writing this thread, but sadly I did not.
