I created a folder named “10.01 Topic,” and later realized I need to split it into multiple folders such as “10.01 Main Topic,” “10.02 Related Stuff to Topic,” etc. However, the IDs 10.02 and beyond are already assigned to other folders. To keep related items together, it would be beneficial to choose approximate IDs rather than continuing from, for example, 10.21.
My options seem to be:
Renaming all folders and assigning new IDs.
Deepening the structure to 10.01.01, 10.01.02, etc.
Both options are not ideal. How should I handle this situation according to the Johnny Decimal system? Any advice would be appreciated.
I think a strict interpretation of JD would have it that the order of IDs doesn’t matter and not to try and introduce structure at the ID level. However, I do understand the lure of having like items together and I’ve definitely been guilty of renaming IDs through my system to follow my latest scheme. This is fairly straightforward in initial phases of a system but gets increasingly hard as time goes on.
The other thing to consider is the newer concept of aiming to expand and generalise an ID to contain more content. If 10.01 and 10.02 are closely related can they happily reside together in one ID? You can have subfolders at the ID level and sub-categories (10.01.01 etc.) probably aren’t necessary for most situations.
When I first started to implement JD, I came across some fascinating documents from the US War Department (this is WWII vintage) about their numerical filing system. Much more digits. But what is fascinating is that they also had an alphabetical “reverse” index to look up the ID numbers. At some point, with complex systems, this may be of use, although perhaps it has been deprecated in the era of efficient search engines.
well, you could use a numbering scheme which allows adding more IDs in between existing ones. Kind of like your option 2, except not making subfolders (if that’s what you meant by ‘deepening the structure’) but adding new ID folders which will sort where you want them to.
10.01
10.02
10.03
add an ID between 01 and 02:
10.01
10.011 (or 10.01a)
10.02
10.03
You can add as many as you want by lengthening the decimal sequences.
Of course, then you have something more like a Dewey Decimal system, not a Johnny Decimal system. If these topics are such that their adjacency really feels important, maybe there could be a system inside a system, i.e. maybe see if IDs can be broad enough to contain a whole topic, and in that ID you could use something like this.
That’s like how people use date-based subfolders where it makes sense, or how I have all my literature references with their own ID scheme in one JDID folder.
To paraphrase the other replies, the ‘general direction’ of JD is towards ID folders that aren’t susceptible to this urge.
In paragraph titled Avoid creating subfolders, it states: If it ever feels like it’s getting out of hand, split the thing in to multiple IDs.
This last statement made me pause. Splitting into multiple IDs (at a later point in time) might lead to IDs that are far from the original one, creating a new issue. Another alternative to the already suggested ones could be to empty the original folder, mark it as such, and then create new folders using the last available IDs.
However, I do understand that IDs should be defined in such a way to avoid this problem in the first place. I am primarily concerned with those rare cases where the number of subfolders gets out of hand.
The adding a suffix (numerical or verbal) to the tail of folder’s name is my way too:
It is a way of a civil engineering standart (GOST) ГОСТ 21.101-2020:
when I wrote my suggestion I was thinking: you could do this but maybe you shouldn’t want to. But thinking on it, I actually feel this might be a great solution for exceptions, as you say, where it make sense that related things are close together.
