Now that I think of it, I do folder in different ways…
For work (lawyer) I use a folder for each client, with inside a folder for each case and inside of it a set of predetermined folders (correspondence, my docs, other parties docs and so on). Each of this folder may have some sub folders if appropriate.
For admin stuff I have a folder for categories (that make sense for me, of course. Apple, insurances, bank etc…), maybe subcategories (in apple there’s a subfolder for every device) and if needed a folder for each year.
For photos I have a folder for year, month and day of shoot; I create other relevant association (summer '21 trip, xyz wedding and so on) with albums and smart albums in capture one.
Many years ago Macsparky did a podcast on this topic and it was around the idea you were either a “hunter” or “farmer”, … the farmer being a very structured folder system. I think the evolution of search, … Spotlight and other search tools, however you choose to do this, makes this hunter versus farmer less of an operational issue than one of taste. I have found with a good naming protocol, its very rare that I cannot locate what I want just based on file name search. I think its useful to split documents into areas of your life like personal, work community, hobbies etc, but beyond that all my documents go into one of 4 folders determined by if I want to retain the document “forever” or at the other extreme 120 days. So for example I have a 7 year folder and have Hazel rule that says if I have not looked at the document in last 12 months and its over 7 years old it can go to the trash. Of course storage is so inexpensive you could keep everything.
I now have everything in DEVONthink 3 and hardly any folders and only three databases, two of which are hardly used now, mostly for personal records and past work of a different type than what I now do. I only have a few folders, some smart ones. I only go one folder deep and the folders, though they make a kind of sense to me mostly wouldn’t to anybody else and are pretty ad hoc, some authors of papers have their own folders, others don’t.
I rely pretty much on the search and AI features and just ‘look see’. Again my needs might not be the same as someone with large client/student bases. Or who needed clear ring fencing for privacy or other purposes. None of that is needed by me these days and a little bit of muddling of boundaries doesn’t matter at all.
I think I use folders in the “conventional” sense, in that I have a folder called Documents, with broad areas of my life as folders inside that, and subcategories as folders inside of those “area” folders, and so on. Don’t really have a whole schema as to when to create subfolders, etc, I just make more and shift files around as the need arises.
Folders are where I put things that I want to search in vain for at a later date.
All seriousness aside, I absolutely love the APFS file system. Thanks to its copy-on-write functionality, I can make a nearly instant copy of a half-terabyte folder and perform a different analysis, edits, etc. on the items within, without taking up much more space.
My structure itself will probably bore folks. I don’t keep much in Documents, pretty much everything can be a “document”, and I sync with ResilioSync, so I have a few other folders set up to sync.
I have a folder ~/Databases with folders inside like Personal, then folders within like Inventory where there are PDFs documenting my autograph collection. Taxes, Health, etc.
I have a pair of external 2TiB SSDs as a RAID0 array where all the interesting stuff goes. …/Research/aim1(aim2, etc.) …/Research/fmri where I brainstorm and learn analysis techniques (with the aforementioned folder copies).
For each class I TA: …/School/Neuro/Neurxxxx/Spr22 (Fall21, etc.)
Also a catch-all folder …/BigStuff for misc. oversized items.
All pretty standard stuff.
Some folder trees get deep, and past me had interesting opinions about where things should be (e.g. I have code, data, and stats folders, but stats has code, data, and stats files within). So I’ve been documenting everything in Curio with links to relevant files and folders, connecting lines for data flow and who is calling whom, etc. It’s a tremendous help with keeping up with things, and well worth the time spent documenting my pipelines.
Then there’s HoudahSpot for finding lost sheep, and I’m getting my tags cleaned up so I can make better use of them.
I have a question about this point. I’ve read many commentators who have made the same point. It gets applied to files, e-mail archives, and even the internet (thank you Google). As someone who is a farmer, by your definition, and who spends a whole lot of time farming, the freedom and ease of being a hunter sounds like some kind of mythical land to me. But it never works when I try it.
So, here is my problem and my question for you. Hunting seems perfectly easy and works when looking for a file or two. Find the monthly accounting summary spreadsheet. Easy. But if I have a project that has dozens of documents, how do I get hold of them all at once. Putting them in folders seems the natural way to make sure they remain glued together. What if I have emails from possibly a dozen people on a project and need to look at them as a group? Seems like folders are the only way to accomplish this.
I hope none of this comment comes across as trolling. I have no issue with the hunters, other than that I wish I could become one. I am on a perpetual quest to find ways to make the computer take some of the burden of mundane organizational tasks away from me – thank you Hazel for being the best automated co-fisherman.
As someone who is also a “farmer”, I totally agree with the feelings outlined in your post.
I suppose eventually software that creates bi-directional links between files, allowing you to link related ones together, will be developed to solve these issues. Pretty sure Hook already does this, although I’ve never used it so I could be wrong. Unigraph (mentioned in another thread on this forum) is shaping up to do something similar.
For work I have the following setup (I only started there in 2020)
When I start a new project, I create a folder in the Active Folder with a title YYYYMMDD - Project name
I work on the item in that folder. When it’s completed, it’s moved into the folder for the year it’s labelled with. If it get’s put on hold or abandoned it goes into the respective folder.
Many projects I work on take weeks or months with dead time due to waiting for others, but I know what’s on the go by moving through my Active Folder on a Friday as part of my Weekly Review, so ensure that nothing falls through the cracks.
I think of folders as fences around files – corrals, perhaps. If I get a file, or make a file, and think I’ll be getting more files similar to it (same topic, same sender, same something else), then I’ll put it into a relevant folder. Or create a folder that will be the corral for that subject. If I will not be getting or making any similar files, I’ll stuff the file into a generic holding bin folder. Or, I’ll drop the one-off file into the desktop, which I periodically clear out.
Thinking of files that way, the files I have in my various storage locations (Mac, Synology, external discs, clouds) are located in folders that grew up over a long time based on judgments I made about similarity of files and likelihood of getting more related files.
I didn’t’ start with a plan – I have no plan. My hierarchy is very workable for me, because it grew up like a coral reef I am very familiar with – one file at a time.
Ah, the Documents folder - that is the last place I will put anything on Windows or Mac OS.
Expected - a place where I can put my documents, arranged as I see fit.
Actual - many apps see it as a place they can fill up with their own folders and application support type information. So it’s not a place that I get to control and keep the way I want it.
So, OS and app providers - if you really want me to start using the Documents folder, you must stop. Otherwise I will carry on using the Desktop :-). This is not some mortal sin, it’s just a place in the folder system where I happen to have a preferred level of control.
As a real farmer it only makes sense that I am a farmer in my data systems too.
I hope you have good backups and run the new check file integrity check on those databases really often. I just had another major data loss in DT3. All indexed files too, none actually imported into DT at all.
Yep that makes sense to me.
I’'m moving a lot of my glue whch was notes etc. Into Obsidian and I am using notes of links a lot like folders for stuff there. That’s fine for the things I want in Obsidian but I still use my full file system of folders as described for most of the the support materials and most of the documents.
That’s why my personal filing system really starts at the folder in Documents called DWA_Data. In there I am in complete control.
Hm. Looking forward to Brisbane’s answer, but until then— two thoughts:
nothing wrong with doing a bit of both, I think: hierarchical folder system for groupings, search for being able to quickly locate specific files…
surfacing groups of documents depends on a common item for the search item. So for example, if you wanted to find all files for a specific project, and all those files had the project name in the title (or somewhere in the text/content) searching for that project name should surface all the relevant files. You could then drill down to find PDFs for that project, etc.
Email’s a whole other thing. I don’t file email in folders any more (I used to do so almost religiously until I figured out that for the way I work it required much more energy than I really needed to expend). Search, for me, works fine, although it is a shame that most people I correspond with don’t appreciate the value/importance of an appropriate email subject line…
I agree with all of this. I’m a farmer when it comes to files, and I’ve always been jealous by hunters who can simply search to find what they want. I’ve tried, but it just doesn’t seem to work for me. I tend to use a very consistent naming system for my projects (similar to what Stephen Wolfram does, and probably influenced by my time working for him), and so my keyword searches can return a lot of similary-named (or sometimes identically-named) items that reside in different subfolders. It’s a little better when my projects have unique identifiers (e.g., Jira ticket numbers) that I can embed into the folder names.
Curiously, I realize that I treat email differently. I have a small amount of folders, but everything else goes into a single folder named Archive. I do a few tricks to make searching easier, like adding project identifiers surrounded by brackets (e.g., [RPM25]) in subject lines, and I also use a very standardized set of prepended labels in my subject lines, which helps me find certain types of conversations that I’ve started.
YES!! If people could get on board with using meaningful subject lines, and when using a year-old message as the source of a new thread, changing that subject line to something appropriate, a huge chunk of my e-mail organization would be wiped away.
I have to use folders to associate emails to case files in my work, so we have an archive in the file system (not just my e-mail repository). This summer I tried the Stack method and just archived everything for work in one flat archive. It felt good, saved a ton of time, but ultimately did not work.
But for personal, non-work projects, and the like, I also use a basically flat archive with only a few specialized folders.
I occasionally use folders to include info next to a folder say of photos “Nice Christmas photos '21”, referring to the folders there or something similar so I don’t have to add more info, rather than adding to the comments box. Easier to see. And certainly not the way it was intended to be used.