A small system

Before adding another tag, I decide how I will search

A project, a topic and a status are three different reasons to label a note.

Stacks of paper documents and file folders.
Photo by Wesley Tingey · Unsplash License. Illustrative photograph.
Contents

It is easy to give a note five labels while being unable to explain how any of them will help later. I would rather start with a search I expect to make. “Show me the material I still need to review for this project” gives a label a job. “Make this collection look organized” is harder to act on.

Here is a constructed library for a short essay about shared neighborhood spaces. It includes a policy page, an interview and several reading notes. I will use it to separate three things that often get mixed together: which project an item belongs to, what it concerns and what needs to happen next.

Start from an actual retrieval question

For the demonstration essay, I would list three questions: which items belong to the essay, which discuss access and which still need checking? Each question suggests a different relationship. A single label called “important” answers none of them clearly. It may describe my mood when I saved the item rather than a useful property.

I would test the names on a few existing items before adding them to everything. If I cannot decide whether a note is “access” or “availability,” perhaps I need a definition, perhaps both terms are useful, or perhaps one should disappear. That small trial is cheaper than maintaining an elaborate vocabulary I do not understand.

Use a project group for project membership

Zotero's collections allow a source item to belong to multiple collections without duplicating its underlying record. For that application, I could keep the policy item in the essay collection and a separate neighborhood project. I would not create another source record just because the same page helps with two questions.

That behavior should not be assumed in every program. A filesystem folder, a note app's folder and a reference manager's collection can mean different things. Before moving anything valuable, I would check whether I am moving a file, changing membership, copying a record or deleting it. Similar-looking interfaces do not establish identical rules.

Reserve a subject label for a useful theme

In this example, “access” would cover material about who can use a space and under what conditions. I would write that definition beside the label list. A note about a building's paint color would not acquire the tag simply because the building is shared. The subject should describe the relevant content, not the whole world around it.

I would also allow a note to remain untagged when I do not yet know its value. An inbox or review list can hold uncertainty without making it a permanent taxonomy. Creating a new subject every time I encounter an unfamiliar phrase may make capture feel productive while making future retrieval harder to predict.

Give status its own small vocabulary

A status should tell me what happens next. For this demonstration, I would try “to-check,” “usable” and “parked.” These are local choices, not official workflow states or a universal classification. “Usable” means usable for the current essay under its stated scope, not that a source has become unquestionably correct.

Obsidian properties provide structured fields with documented types. If I used a status property there, I would keep its meaning consistent and follow the application's format rules. In a notebook, a simple status column can do the same conceptual job. The field is useful only if I change it when the work changes.

A tiny label register
Project: neighborhood-essay
Topic: access — who can use the space, and under which conditions
Status: to-check — one stated question must be resolved
Status: usable — enough support for this project’s stated claim
Status: parked — retained, but no current action

Run the retrieval test before expanding

I would now ask the original question: can I find the essay's access material that still needs checking? If the tool supports combined filters, I can use those distinctions together. If it does not, a short project index may be simpler. I do not need a new application just to justify a labeling scheme.

The test should include an awkward item. Suppose an interview concerns access but has no pending check because I decided not to use it. It belongs to the subject without belonging to the active task list. If my labels cannot express that difference, I would revise them before expanding the library.

Retire a label with care

Labels can outlive the questions that created them. I would review a small set when the project closes: which helped, which overlapped and which meant different things on different days? A short note about the retired terms can preserve the context of older work. A complete relabeling exercise is not always necessary.

Zotero distinguishes removing an item from a collection from moving it to Trash. That is exactly the sort of distinction I would verify before a cleanup. I want to remove an unnecessary relationship without accidentally discarding the record or losing the only source pointer behind an old claim.

Build a small library before designing its labels

I would test the idea on a deliberately small collection: a room-booking policy, an interview transcript, a meeting note, a draft outline and a useful reference about public spaces. These are imaginary records for the neighborhood essay. I would put them in front of the actual retrieval question: which material about access still needs checking? The policy and interview may concern access, but that does not mean they have the same next action. The meeting note may contain a decision rather than source evidence. A single “important” tag would blur all three distinctions.

For the policy, I might record that the question is whether its wording applies to the proposed activity. For the interview, I might need to check attribution and permission for a particular use. For the outline, there may be no external fact to verify; it is simply the place where the argument is taking shape. Those differences suggest a status field or task note, but only after I have described the work. I would not create a complicated status vocabulary and then search for differences that justify using every term.

Next I would perform the search manually on this sample. I would choose the essay's project group, identify access-related material and look for a stated unresolved question. That exercise tells me what a filter must preserve. If the tool cannot combine the exact criteria conveniently, a short index with links may be clearer. I would not switch software before checking whether the collection is small enough for that simpler solution. The organization should fit the work's scale and the people who will maintain it.

Finally, I would test a misleading match. The interview might mention the word “access” in a different sense, such as access to a digital document. A text search can retrieve it without establishing that it concerns building use. I would label the relevant concept carefully and inspect results rather than treat the presence of a keyword as successful classification. The test is whether I can find and interpret the intended material, including recognizing an apparently relevant item that belongs to another question.

Separate topic, status and the next action

A topic is usually a statement about content. A status is a statement about the work's current state. A next action explains what someone should do. I would keep those roles visible even if the application stores them in the same general labeling interface. “Access” can describe the policy's subject. “To check” can describe its state. “Ask whether the event falls within this clause” describes an action. The third cannot be recovered reliably from the first two unless I record it somewhere.

This matters when the work changes. Once the clause is clarified, the source may remain about access while its status changes. The action can close without removing the subject label. If all three meanings live in a single tag, a cleanup becomes ambiguous: am I saying the item no longer concerns access, that it is no longer active, or that a question was resolved? I would prefer a short explanatory note to inventing a combined tag for every possible state of the project.

I would also define what “usable” means locally. For this essay it might mean that I have inspected the relevant passage and know the limit of the claim I intend to make. It should not mean that the source is universally correct or that every statement in it has been verified. A status needs enough scope to remain honest. If the same record later supports a different claim, I may need another check even though the old project's question was resolved.

The action should name the smallest useful uncertainty. “Research policy” is broad. “Confirm whether the proposed session counts as an open public event under the current booking terms” is much clearer in the invented example. I would keep that wording with the project note, while retaining the policy as a source record. This avoids turning a temporary task into a permanent subject category. It also helps another contributor understand why the source is waiting, rather than assuming a generic review label means it has been neglected.

Test the naming system on a completed project

I would imagine the essay finished and ask which labels still serve a purpose. The project group may preserve what informed the final document. The subject label can remain useful for future reading. The status “to check” should not persist simply because nobody reviewed the labels after submission. I would inspect the actual unresolved question before removing it. The work may have ended because the claim was dropped, which is different from establishing that the source supports it.

That difference belongs in a closure note. “Not used; question no longer needed for this essay” preserves the reason without pretending verification occurred. If another project later retrieves the source, it can begin from the actual state of knowledge. I would avoid cleaning a finished collection by marking everything complete at once. A tidy status field can conceal unfinished factual work, and the next writer may interpret completion as stronger assurance than the original project ever provided.

I would check label meanings with a new reader too. If they interpret “review” as an article type while I mean an action, I can choose a clearer term or separate field. I would not solve every ambiguity with an abbreviation. A smaller vocabulary expressed in ordinary words can be easier to hand over than a code system that requires a separate legend for routine use. The legend should explain genuine distinctions, not compensate for names that nobody would naturally understand.

A finished project also provides a natural opportunity to retire unused labels. I would inspect the items affected and note the change if historical searches depend on it. Removing an empty or redundant topic label is different from deleting records. The application's documented behavior matters here. I would make one small reversible change and confirm its result before undertaking a larger cleanup. That keeps a useful refinement from turning into an accidental loss of the source context I was trying to preserve.

Know when a label should become a note instead

Some categories grow because they contain unresolved thinking rather than a stable topic. Suppose I keep creating “access-but-not-evening,” “access-exception” and “access-unclear.” I would ask whether those distinctions belong in an analytical note about the policy. The note can explain conditions and compare passages; a chain of tags may merely encode fragments of the argument. Turning the comparison into prose can make the reasoning easier to challenge and revise without changing the classification of every source.

I would try writing the comparison in three sentences. The first states the apparent rule, with its source pointer. The second describes the exception or uncertainty. The third identifies what the project needs to decide. If that explanation solves the retrieval problem, I may need only the broad subject label and a link to the analysis. The source collection and the argument can cooperate without requiring the taxonomy to express every turn in the writer's thinking.

This is especially useful when several people contribute. One person may use a tag to mean a theme, another a worry and another a reminder. I would ask contributors to explain their intended retrieval question before standardizing the vocabulary. A shared definition can preserve good distinctions. It can also reveal that two similar labels serve different legitimate jobs. Merging them solely because their names look close may make the collection shorter while making the original questions harder to recover.

My stopping point would be a successful search and an understandable record, not a theoretically complete subject map. I would want the active essay material to be findable, the remaining checks to be explicit and the meaning of the few labels to be clear. Other parts of the library can wait until a real task exposes a problem. That boundary keeps digital organization connected to the work it supports, rather than allowing the arrangement itself to become an unlimited project.

Test one search before adding the next label

I would finish by writing a search request in ordinary language and answering it with the sample library. Can I find the essay's access material that still has an unresolved source question? Can I explain why each result appears? A misleading match deserves inspection, not another label added automatically to compensate for it.

If the answer is clear, I would stop expanding the vocabulary. If it is not, I would identify whether the problem is a missing definition, an unclear action or the application's retrieval behavior. Those are different repairs to the system. A useful label collection remains accountable to the search it supports, rather than growing because every unfamiliar phrase appears to deserve its own permanent category.

Frequently Asked Questions

Q: How many tags should a note have?

A: There is no useful universal number. I would use only the labels that answer retrieval questions for that collection. A source may need several subjects, while a simple task note may need none.

Q: Should I use folders or tags?

A: Begin with the behavior you need and the application you use. A folder can provide a clear home, while a tag can connect themes across groups. Check what moving or removing an item actually does in your tool.

Q: Can I use the same word for topic and status?

A: You can, but ambiguity makes searching harder. I would distinguish a source about “review methods” from an item whose next action is “review.” A short prefix or separate field may be enough.

Q: What should I do with old inconsistent tags?

A: Start with a small, active part of the collection. Document a few preferred terms and reconcile genuine duplicates. Preserve useful context rather than deleting unfamiliar labels in a single sweep.

Q: Will a better taxonomy make my research reliable?

A: It can make material easier to retrieve, but it does not verify a claim. You still need to inspect sources, preserve attribution and decide whether the material supports the particular use you have in mind.

One thing to take forward

Before adding the next label, I would finish one sentence: “I will use this to find…” If the ending is vague, the label can wait. A small vocabulary with clear jobs is easier to maintain than a comprehensive-looking one with no retrieval plan.

Sources & further reading

Something needs correcting? Write to the editors with the article link.