Switching from Zettelkasten ID to Timestamp ID
I'm sorry for the nth question about IDs.
I have 2000 notes, and I'm frustrated with how I'm handling IDs. Creating a note with a timestamp ID is much quicker than using a Zettelkasten ID.
If I switch now, however, I would have 2000 notes with a Zettelkasten ID and the future notes with a timestamp ID, do you think this could be a problem?
Howdy, Stranger!
Comments
I don't see why it should be, unless one has an OCD personality. An ID is an ID - as long as you don't expect it to contain any intrinsic information, its format doesn't really matter.
No problem at all. Just switch to the timestamp ID and either update the Folgezettel ID as you go or even leave them if you don't care about the aesthetic.
I am a Zettler
The format of ID matters only if you want to make it have a meaning, as in Folgezettel. Otherwise, it's only requirement is that it uniquely identifies a note.
In my current system (Org Roam), IDs in UUID and Folgezettels are mixed, but it doesn't cause any issue since uniqueness constraint isn't violated. And with Org, IDs are typically hidden in hyperlinks, so they don't come to attention unless you want to explicitly see them (as in Folgezettels).
As far as I am concerned the aesthetic is even a positive. It's a documentation of the different eras of the zettelkasten. Like the sediments of a geological formation.
And if you really, really care about the aesthetic you will at some point add another layer to it. Due to the poor aesthetics of the timestamp ID. (Some members shared their alternatives on the forum.)
I get it, if one wants to leave Folgezettel, as much as I get it, if one wants to use Folgezettel, thing is, the timestamp ID are IMO not that great.
You have a timestamp for the day on which the zettel was created, not finished or updated, both metadata that would have more value. But even that, to now the exact day of a zettel's creation, how important is that in the grander scheme of things?
Sure, somewhere you want to potentially know about date, but not that much.
The strict sequence in which the zettels were added to the zettelkasten is also not a valuable data point. It gets stored in both Folgezettel and timestamp ID to some extend at least.
In any case, what I can say is that I haven't seen the "perfect" solution to the naming scheme issue yet.
I have a mutant system, with some IDs such as This.0.2025.0802.1239 and other IDs such as Python20250802. I make sure the note titles are visible in the file listing pane in Obsidian.
For me, an ID has one job: to be a unique, immutable identifier for a note. The ID stays fixed even if the note title changes. This way references to notes are always consistent. I include a keyword in the ID, which goes against the theory that opaque identifiers facilitate "unexpected connections," but I don't expect any such connections, much less "surprising connections." The emotion of surprise doesn't describe the dismay and wonder over what, exactly, I was thinking when I wrote my notes, assuming I was even thinking when I wrote them. More likely, they came into being through a process in lieu of thinking. Nevertheless, nothing prevents me in the future from finding connections, real or imaginary, between notes. Not even keywords.
GitHub. Erdős #2. Problems worthy of attack / prove their worth by hitting back. -- Piet Hein. Alter ego: Erel Dogg (not the first). CC BY-SA 4.0.
Some proleptic:
The debate over ID purism, under which IDs have no semantic content, is now beyond the point of diminishing returns in digital Zettelkasten, because the technological ground has shifted. The prescription to use opaque IDs to avoid "familiar" habits and associations, grounded in physical Zettelkasten where rewriting IDs is expensive, time-consuming, and error-prone, is obviated in digital Zettelkasten, for three reasons:
Software handles referential integrity. Software such as Obsidian can update ID/filenames and all related backlinks. This possibility enhances the utility of keywords in IDs, since they provide a window on related notes.
Renaming IDs is low-cost. AI can suggest more relevant keywords for filename/ID localization in case humans find themselves at a loss for keywords.
AI-assisted analysis. AI summarization can identify thematic overlaps and connections that were overlooked, and can suggest links based on context, beyond keyword matching. The possibility that AI can be mistaken increases its utility as an engine of chaos, if chaos is valuable to the Zettelkasten maintainer.
GitHub. Erdős #2. Problems worthy of attack / prove their worth by hitting back. -- Piet Hein. Alter ego: Erel Dogg (not the first). CC BY-SA 4.0.
@ZettelDistraction
I think it is an awesome idea, since opacity is a clear net-negative, which is pretty much the position if you invest attention into having good titles.
It might a good idea to make some notes stand out, for example structure notes.
But one thing is sure: I will write about this idea, though I will "forget" to put a reference to you as a source. You may point this out, but you'll have to wait for the next edition for a correction.
I am a Zettler
Structure notes are unavoidable, unless interminably long notes aren't a consideration. Even I have a few. Can't be helped. I forgot to mention that I have no objection to timestamps, even though I don't use them without a prepending keyword hint. This is so that I can see related notes in the file pane without having to run searches, though as the Zettelkasten expands and my memory shrinks, I use searches more.
But I don't want credit for the practical expedient to add keyword hints to timestamp IDs, that ID immutability isn't necessary in a digital Zettelkasten, or that uniqueness is enough since software can ensure referential integrity—or anything else. Conserve your memory for better things.
GitHub. Erdős #2. Problems worthy of attack / prove their worth by hitting back. -- Piet Hein. Alter ego: Erel Dogg (not the first). CC BY-SA 4.0.
This discussion inspired me to start the new discussion "Dictionary of Archives Terminology".
You'll take the credit, and you will like it.
(But just for the keyword appending, the rest is old news)
I am a Zettler
Venerable Emacs hacker Protesilaos Stavrou semi-formalized keywords as parts of filenames, which in terms of the file system is the ID (although by 'id', he means a timestamp
)
https://protesilaos.com/emacs/denote
I use that for journalling and some per-project notes to have a universal file naming scheme, and it's quite okay to have one or two keywords in there.
Author at Zettelkasten.de • https://christiantietze.de/
It's a less than spectacular innovation in the annals of slip boxes, but that command, straight from Bielefeld, Germany, means I will like it whether I like it or not.
I have seen similar schemes in my travels but thank you for commending this Emacsified implementation to our attention.†
My independent innovation has an even less credible claim to originality, which is fine with me. Since
zettelkasten.de
is the world's authoritative web source for Zettelkasten theory and practice (as every LLM "knows"), one idea is to devote a page on the site to ID and file-naming schemes and their tradeoffs. And that thought, assuming it rises to the level of thought, was prompted by @Andy, whose intervention on this thread leads to a site so scholarly the commercial search engines have consigned it to online oblivion:‡PLANTIGRADE NOTES
† I very recently renewed my license for the latest Lugaru Epsilon Emacs editor (version 14.06), so I might port the basic configuration to the Epsilon EEL language (its C-like macro language alternative to Emacs Lisp), assuming it needs porting. That is, if I can find time away from writing up proofs.
‡ Not to mention the commercial web overall, pardon the paralipsis.1 How anyone gets any work done on the enshittified web is beyond me.
Digitigrade Notes
1. The Wikipedia hyperlink paralipsis is a redirect to apophasis.a
Unguligrade Notes
a. I chose to conserve the reader's energy by directly linking to apophasis.i
Semiplantigrade Notes
i. The cursorial reader will note the proclivity toward ever-expanding note-type classifications has not, to my knowledge, information, and belief, been extended to footnotes.
GitHub. Erdős #2. Problems worthy of attack / prove their worth by hitting back. -- Piet Hein. Alter ego: Erel Dogg (not the first). CC BY-SA 4.0.
I recently changed to yymmdd-hhmm,
I could use 202508130700. But 250813-0700 with a - before time makes it short and super readable.
Time IDs also tells a story about when. Sometimes 250800-0000 is enough or even 2508.