@webwalrus I agree; there is a limitation in the JD system by only allowing 10 categories in a given area. Further, the JD website does acknowledge that rules may need to be broken at times, eg creating a third level of folders in some cases.
This is part of why I was thinking about how I would have three digits or a letter and two digits for categories, because again while I think 10 areas is enough for me, I clearly will have some areas with more than 10 categories, as I already have folders in my current scheme with more than 10 subfolders.
If you have 50 active clients at a time, however, then I am going to guess you have more than 100 clients total, so even 3 digits is not enough. You would really have no alternative but to have, for example, an area for Work with categories 10-19, and then perhaps your folder is ā11 Clientsā, and under that you would have a folder for each client and under that subfolders for the different assets for the website of each client.
Clearly the āstrictā JD system would not work for you.
I agree that as a folder fills with hundreds (thousands?) of files, you replace drilling down into a folder tree with scrolling. In your situation, I would tag the folder for each active client with an āactiveā tag and use a SmartFolder to quickly access the active clients as I described above. Once I was in the ā11 Clientsā folder faced with an array of 1000 or more client subfolders I would use the search field to quickly find the one I want.
If you have 1000 clients all represented by a folder in a Clients folder, you are going to have to scroll regardless of whether you have JD in place or not!
Thatās interesting, because their short summary explicitly says thatās not allowed. I find myself wishing people advocating these systems would just say what they mean.
Currently I have an āArchiveā folder where I periodically shuffle clients folders that I havenāt touched for an extended period of time. Thereās no point in them being in the main process flow most of the time.
But in each of those folders, for example, Iāll frequently have multiple sub-folders broken down by project. Even if I break the rules and allow more than 10 items in a sub-folder, itās just not as simple as āonly have a single-level folder for each clientā unless my filenames are something like:
Basically Iād have what would otherwise be folder labels contained in the filename. And that just doesnāt make sense to me, especially when it creates a ginormous folder with 10,000 files instead of several well-filed sub-folders.
I was reading a little about Johnny Decimal this weekend. I get the impression that thereās a level of folder structure that heās not talking about, which is the project level. That level is above the top level of Johnny Decimal.
Letās say you work on the production staff of an episodic TV series like The Simpsons. Each episode has a number of required steps and reference assets. Those steps and types of assets are the same from one episode to the next. Each episode gets its own folder, and within that folder you organize all the steps and reference assets for each episode using the Johnny Decimal method.
But you canāt really organize the episodes themselves using Johnny Decimal because it doesnāt make sense and because they have been 734 episodes of The Simpsons. And thatās more than 10.
I immediately thought of creative agencies as well.
ABC Corp
22014 2022 media buys
23111 2023 Website updates
23147 Spring campaign
01 Briefs
02 Supporting material
03 Project files
Banners
Copywriting
Creative for web
Print media
etc. etc.
Usually with some kind of automator script to spit out the folder structure for each project. Any changes to the structure are announced in emails with all cap subject lines with a hard cutoff date.
YearClientCodeSequentialNumber is also a good structure if you can support two or three letters for the client codes (otherwise all the numbers are too hard to read.) E.g. 23ABC111
I tried to implement this and immediately got lost.
Say 1 is Learning and 2 is Finance, I now need to remember that. If I just called the folders Learning and Finance, theyāre a lot easier to find because theyāre in some sort of order (alphabetical).
Iāve found myself using the JD inspiration of large buckets with categories underneath, but the numbers make things harder to find so Iāve got rid of them.
Now think what it would take to train 150 people to use the system.
J.D.: āYou can search for things, but the results are garbage.ā
Iāve been using search for the last 15 years or so. First with Evernote, now with EagleFiler for local files, and Google for my online files. So far, so good.
And if half the claims made by Microsoft, Google, and others are correct, it is only going to get better.
I think itās easy to go overboard with JD notation and create a complicated coding scheme that nobody would use. But I have also found that for personal usage the numbers acts as a some short of mnemonic cache. If I happen to be doing my taxes I will surely remember the code and can jump to the folder right away with Raycast or something similar just by typing the prefix instead of navigating the folder hierarchy. Thatās pretty useful.
I think thatās really the key, and filing systems get too complicated because they get overengineered.
This is a David Allen tip from one of the GTD books. Basically stuff for your 2015 Cadillac Escalade should be filed in the simplest category you can think of. āCarā or āVehicleā work well. Maybe āCadillacā. But ā2015ā is probably going to be less likely. ā2016ā (the year you bought the car) is even less likely. And āJakeās Oil Changeā is almost certainly not going to cross your mind when looking for vehicle records.
Just putting everything car-related in a folder called āCarā, with the stuff in there going roughly from oldest to newest front to back, makes things much, much more findable.
I donāt run JD in my system, but I do have numerical prefixes in my personal file structure and it works fine. Iāve been running it for several years and for folders I use regularly I know what number they are. In any case, not everyone wants their folders arranged alphabetically. Mine are arranged in order of priority/use.
Also, working for an organisation that currently has no file structure (actually worse, a file structure where each employee is implementing their own system in their own little kingdoms), I can tell you that search isnāt the answer itās heralded to be. Sometimes you donāt know what youāre looking for, you just want to go in a folder for a project and see whatās related files there are. Search canāt do that. (Itās currently impossible in the mess Iāve mentioned, which means you have to hope youāve used a good keyword and that the search results are picking up anything that might be of relevanceā¦. I know it isnāt!)
Building, implementing and encouraging use of a new structure is in my workplan. Of the colleagues Iāve polled who have an opinion on these things (many donāt, hence the current mess!), most favour a numerically-led system. As I do too, thatās what I will be presenting as the best system for our needs. Again, this is a YMMV situation, obviously if my colleagues all wanted alphabetised folders, or dated folders, weād go with that (we are in work where dates arenāt relevant for top level file management).
Incidentally, I also arrange my Apple playlists numerically! All have a 0X.XX prefix then their name, where the first two numbers indicate the type of playlist it is (01-09), and the second two numbers indicate which list it is in the ācollectionā. This means that themes are nicely grouped together. E.g. all my single artist playlists are 02. All my Christmas lists are 09 so theyāre at the bottom out of the way. Etc.
My first job after I left school was as a filing clerk for a major oil company. This was in the days of paper filing and these huge rooms with rolling cabinets. We even had a training course for filing - one of the things I remember most vividly was the that the starting point for a filing system is to think about how you are you going to retrieve information, not how you should file it.
I am downloading the episode ā¦ and following the podcast because why not?
I dug deep on Johnny, decimal one weekend and tried it for a bit. I quickly decided it was not for me, but my current organizational system has elements strongly inspired by it.
Thanks for the podcast, listening to the discussion had me consider using Johnny Decimal, but after thinking about it and reading the website, Iāve decided I donāt need it. I have an organized system based on folders and tags. Iāve been using it for years now and I rarely have trouble finding things. I do like the idea of using numbers to categorize things, it appeals to my nerd side. But if I can find things now, not sure it is worth restructuring my system.
I enjoyed the discussion about work. All that is very true, people canāt find anything at work. I keep my stuff organized and people sometimes comment on how I can find things, but no one is stepping up to propose a system we would all follow (including myself).
agree with this, especially the overengineering comment.
Itās good to be open to discussion and considering alternate options. However, I find itās best to do what comes naturally. I have to deal with this question quite often, āHow should I structure my database?ā. My answer is always, "Well, how would you naturally do it in the Finder in a way that makes sense and is efficient for you? Okay, now apply that to DEVONthink. Generally speaking, the approaches can correlate.
PS: In your vehicular example, I would be inclined to make the top-level groups 2014 Mazda 6 or 1974 Super Beetle Type 1.
This is the way, when archiving stuff I always think to myself: āWhere would I look for this receipt/manual/whatever?ā And there it goes. My main folders have a numeric prefix (idea I got from JD) so this usually happens quickly.
In my personal vault, I donāt happen to have that many areas of interest, but in my work vault it gets more complicated: some projects or proposals have useful assets that I would like to reuse when I am doing research (like whatās the pricing structure of vendor YYY? what are the business domains of an insurance company?) and then I just liberally duplicate the item to where I think I would also go looking for (or use replicants in DEVON, by the way)
In my experience, one either has a complex folder structure or a complex tag structure, but not both. And tags do not feel very portable to me, so I favor folders. Disk space is cheap and time is not.
agree with this, especially the overengineering comment.