Labels, Tags, and Flags… Oh My!

I wanted to hear everyone’s use cases for labels, tags, and flags. Do you use all of them? Some of them? None of them? Are there benefits to some, and none with others? Does DT3’s AI favor any one of those over any others?

I’m a huge group user, and those make a lot of sense for me. But I’m wondering if there’s a cool way to use labels, tags, or flags that I’m not considering. I don’t really have a huge problem to solve, I’m mostly just curious and would love to hear how everyone is utilizing all these organizational features. :slight_smile:

1 Like

I don’t use tags at all. However the idea is to hold my powder in case I ever really need a cross reference system. For example to separate out some material connected to a particluar author. I have one or two of those, large numbers of papers about and by one author. I am cautious about getting it right from the getgo though and don’t need it yet. I use Groups and some recommend not using both Groups and Tags. I used to use labels. I didn’t find I really needed them, same for flags. I have a vestigal like for “unread” though and sort of use that like a flag. Not an efficient use I will say, but it is my habit.

This is another application:
Tags - cross-reference for grouped documents
Labels - visual indicator of action
Flagged - folders with a certain time based action attached

Labels: I use them for categorizing my work. I have some labels like

  • urgent
  • due this week
  • due some day
  • etc

I created some smart groups that use these labels and so I have a quick overview what I have to do next

Tags: I use them only for all that information stuff I collect in DEVONthink. I collect articles etc. and tag them with keywords. So I have a second (chaotic) classification system additional to the static group structures. But I normally do not use tags for my business documents, here I prefer a reliable group structure

Flags: Do not use them

1 Like

Labels I use only very sparse and actually not really much. But tags I use for everything. I have got a structure of tags with the main (awaiting answer, do be done, etc.) which I have in Finder as well. As I use so many tags, I use a little peace of software called Ammonite by soma-zone which makes searching tags in DEVONthink and in Finder much easier.

Flags I do not use at all but only the „unread“ items.

Hope that helps

1 Like

I use Tags mainly for support nowadays, i.e., not for my own organization.

I use flags on occasion for finding particular documents by adding item:flagged to a search.

Labels are usually just used for testing but also sometimes as a visual indicator when I’m passing some documents around internally.

1 Like

I find that very useful when working through a long list of documents. If I have to stop before completing the list I simply flag the next document so that I can find the right place when I resume work on the list.



Thanks everyone, these are great thoughts.

I’m glad to hear that I’m not missing out on much with labels. I’ve always wondered about them, but I mostly stick to groups and then add tags if I need another level of organization. I find that groups generally satisfies everything I’m looking for the vast majority of the time.

1 Like

We do have a set of labels in-house for databases we sync with each other. They definitely can be handly but obviously not required.

1 Like

I’m still thinking about a way to automatically change labels when moving a replicate out of a folder. Might I “hijack” this topic and ask your input on this Jim?

The use case at hand are multiple personal “pseudo-Inboxes” (i.e. groups in the actual Inbox) that contain replicates of newly imported files with content that is potentially relevant to all users. Users review the content of their pseudo-Inboxes, and either move the file to an appropriate group or trash the replicate if already moved previously by another user.

The idea of automating the label change when moving a file, is that it sets a clear visual marker that the file has already been archived by someone and thus any replicates can be trashed. You end up with one original visually labeled as “Archived”.

I’ve tried this before with SmartRules, but failed to come up with an appropriate trigger/action combination. The database at hand is synced through a sync store on a NAS by the way.

1 Like

At first glance, an On Moving event trigger would be a possibility.
Since you’re syncing, an After Synchronization trigger might do it.

Do realize the label will apply to all instances of the replicated file(s).

Ah, yes I forgot to mention one crucial part. The moves are made “on the go” with DTTG.

The On Moving event trigger appears to be taken quite literal. The file has to move by drag and drop or menu action on the macOS machine running this Smart Rule for it to work. Moving it with DTTG (and thus changing it’s location attribute in DT’s own underlying database I presume) doesn’t seem to be considered a “move” by the Smart Rule.

To put it into other words, if I:

  • put a file in group A
  • label it red
  • sync the database to the NAS sync store
  • sync a DTTG client
  • move the file to group B with DTTG
  • sync the database on DTTG
  • sync the database on DT3 with Smart Rule

…the Smart Rule doesn’t run although the file has “moved” on the DT3 machine.

I’ve also tried the After Synchronization trigger but it seems to be impossible to have a Smart Rule detect egressing movement of a file from a certain group, and act upon that.

Absolutely. That would actually be the point, as it indicates to other users the file has moved to another group.

Since we now have Custom metadata, the need for Labels became a kind of a ghost…

Nevertheless, I can throw in a lifehack of how I still use labels:
You can name, say a Red color label like “Important” (or “Favourites”/ “Currents”) and mark all your important and most frequently used Documents and Groups with this label. Then make a Smart group “Important” (ot whatever) which collects all red-labelled items for you. That’s it. Why it may be interesting:

  1. The most important: you may work with all your selected Groups in parallel. The current drawback of DT3 is that it doesn’t remember the Group’s collapsed/expanded state for each Group, i.e. if you select one group (e.g. in Favourites section of a sidebar) and expand it in needed structure (in according Groups view), DT3 will remember it until you do the same with the other group. If, after it, you choose the previous group, the collapsed/expanded state will be completely lost. It may be very frustrating if you want to simply drag an item from one place to another (like in two panelled file managers)… The only way - is to create this collapsed/expanded state of two or more selected groups in the sidebar, but: (1) you don’t have a horizontal scroll here, so if there are more than 4-5 group levels it becomes unreadable and resizing the sidebar will eat the space from group and document views accordingly; (2) required groups may be in distant locations from each other.
  2. You unpopulate and declutter the sidebar, leaving Favourites section only for groups you want to see in saving dialogs
  3. You can flawlessly award and dismiss the items with this “Important” status, move to another Label, like “On Hold” or smth

I use flags for marking the results of automated actions if it needs a manual check. Like I have a script to OCR PDFs without text using a FineReader application. It flags the result, so I could check the quality and revert to original if needed. Or script which checks the link between DT and Bookends in publications. It regularly updates the Item if you change info in DT or Bookends, based on the double-entry (id of DT item in BE and id of BE item in DT). It can repair this connection if double-entry was broken (e.g. from DT side when you OCRed a PDF - item ID changes), but it flags this item for you to make sure manually if repair was correct.

Since I’m a heavy group user, I use tags only as group tags for selected groups. It helps to quickly and manually classify (replicate) material to needed groups.
I’m currently thinking on using ordinary tags, automating it with content analysing features of DT, but still in a doubt… I think I don’t need tags if I’ll have to manually set them, so I’d accept using ordinary tags if it is done automatically, but still not sure that it’ll be more useful than just manually checking “See Also” or “Concordance”…


The problem I found with labels previously is that they were more cumbersome to add and it sometimes made an extra step: well the way I was doing it, guess I could have found a better way but I never got that far. I think that was the only reason I stopped using them. Fact is you have so many ways of picking things out on this app. I do like color labels though. I always ponder reintroducing them. It would be a big job re labeling old stuff though.

Yes but Smart Rules could make that job a lot easier if you wanted to do it

1 Like

Yeah, I nearly did that but then I realized that if I could find them by Smart Rule then, well I didn’t need the labels. However as you imply sometimes one wants something for idiocyncratic reasons or just to preserve a system one is used to. Especially visually, which is where labels are useful. Thanks for the reminder and I might just do that! Thanks for taking the time. It is worth adding that my filing systems, even before I went paperless, did have some color things I am very used to. Totally idiosyncratic. Certain authors, a very small number had Yellow files associated with them for example and it second nature to me to look for that color.

And some things you do as it seemed to make sense at the time, until at a later time you realize it wasn’t really useful at all! :slight_smile:

For example, my financials database is housed in an encrypted sparsebundle disk image, and is only one of two fully indexed databases I have. (The other database is also on the same disk image.) Looking back at it, there is absolutely no advantage to how I did it and the only reason I can come up with is it was early in my DEVONtech career and I was likely just testing indexing as well as the performance/security/robustness of running the database in the disk image. :stuck_out_tongue:
I leave it as-is, as there’s no harm done, but if I started from scratch I would go all imported like I normally do.

1 Like

Glad to see that last post withdrawn. It’s the second time in this thread alone that you’ve poked fun at users in another thread who care to use DT in a manner different to you.

Although they might have become a ghost to you, I respectfully disagree here.

The benefit of labels really depends on the way one uses them:

  • based on their associated name/text
  • based on their color
  • any other way I can’t think of right now, but might be in use by someone

Another thing to keep in mind, AFAIK the current version of DTTG doesn’t support custom metadata (but I wouldn’t be surprised if that changed in the future).

Labels can be depicted in the DT3 web server but custom metadata cannot.

I am really hoping that @cgrunenberg can someday add custom metadata to the webserver columns - it is a really major limitation currently. For now, labels provide a limited workaround.