devonthink tags and groups

Hi,

I am having trouble understanding the rules of tagging and groups. I currently have the “exclude groups from tagging” switched off. I have read through the extensive explanation on page 95 of “take control of getting started” guide, but I’m still a little fuzzy about how you can have several groups with the same name that act differently. For example, I have made a tag, then created a group with the same name as the tag. I expected the group to automatically have replicated into it any file that has that tag. However, this didn’t happen. Furthermore, when I tagged a new document it didn’t automatically replicate into that group either (although it appeared in the tag section). Can someone explain to me how this works? Thanks,

Jonathon

I suspect that an entire ‘Take Control’ book could be written on tagging in DEVONthink, and it still would not cover all the possibilities. Having said that, here are a couple of quick thoughts to start the discussion.

In my opinion, you do not want to do this. Having multiple tags with the same name is possible (you can even have multiple regular (blue) tags), but it leads to confusion and inconsistent results. You will end up with documents tagged with the same name, but they will not all have the same tag.

Your expectations for how tags should behave is important, as you do have a lot of flexibility as to how to set up tags to get the most from them. Given that:

  1. If you want documents to automatically replicate to a group when a tag is applied, enable tagging for that group and do not create a regular (blue) tag with the same name.

  2. If you want the same tag group to appear in multiple parent groups, replicate the group instead of creating another tag group with the same name.

  3. If you want all documents in a group to have common tags, but you do not want all documents with that tag to appear in the group, then;
    a) create groups that do not have tagging enabled
    b) assign regular (blue) tag(s) to these groups with the info panel for the groups.

    If you do this, note that you can have tags with the same name as groups, unlike the tag strategy above which can result in multiple tags with the same name.

Having said this, you can see that tags can be a powerful, primary tool for classification (Options 1 & 2 above) or a powerful, primary tool for searching and retrieval (Option 3 above). You can also use a combination of all 3 strategies, but I find this to be something of a compromise. Over the years I suspect that I have tried every possible combination of tagging, and Option 3 works the best for me. DEVONthink already has powerful classification capabilities as your database grows (click on the scorcher’s hat to reveal the See Also & Classify pane), so tags as a primary means of classification/replication has less value.

Also, using regular tags lets me create tag names that are short (usually one word) that can be applied to many documents, while group names tend to be lengthy and specific. So instead of having a group tag for a group named “Knowledge Management in K-12 Education”, I would disable tagging for the group, and apply the regular (blue) tags of “KM” and “K12” to this group via the info panel as shown above. Then I can reuse the “KM” tag to groups dealing with knowledge management in secondary education, post-doctorate, and workplace documents and retrieve all my KM documents with a single tag.

Clearly I have my preference for tagging, but that does not mean that other methodologies are not useful-I enjoy seeing how others structure their databases as well. The main thing is that you use a structure that gives you predictable results. Does this help to get you started?

1 Like

@jonathon
you can turn off group tags for the entire database too, not just for a single group. (in the properties of the db)

@greg, could you elaborate on this please:

“Having said this, you can see that tags can be a powerful, primary tool for classification (Options 1 & 2 above) or a powerful, primary tool for searching and retrieval (Option 3 above).”

why are the three different options used for different tasks?
why is option 1/2 mostly used for classification and not search and on the other hand, option 3 apparently for searching/retrieval? I don’t quite why and/or how you would effectively use them for either scenario.

Options 1 and 2 (and Option 2 is really just an extension of Option 1) allow you to classify/file documents into multiple groups without using the AI classification that is already in DEVONthink. One could have documents in the Inbox, ‘classify’ them by applying group tags, then delete what is now a replicant in the Inbox. It’s pretty quick and painless, assuming the database has a small enough set of groups to where the user can remember all the groups where (s)he wants the document replicated. It is also helpful in a small database because the AI does not yet contain enough data to make the “See Also & Classify” function as useful as it can be.

As the database, and the number of group tags, the group tags are not as useful as they once were. The number of group tags in the database must grow, not necessarily to filter documents “in” as it is to filter documents “out”. Using the above example, I must create additional KM in education groups as “Knowledge Management in K-12 Education” is not a fit for KM documents that fall into other education groups. As a result, the group names are now are not as memorable as short, preferably a single word, regular tags.

The above limitation could be overcome, if one’s memory and/or diligence is strong enough to assign the correct group tags. However, I have found the greater shortcoming to be that group tags, when used extensively or exclusively, hide the relationships among the data that metadata is intended to reveal. The group tags really only apply at a primary level. The regular blue tags allow me to assign secondary and tertiary relationships to the data that would be hidden when using only group tags. If I want to retrieve learning assessments for science that apply Bloom’s taxonomy across all learner age groups, I’ll need to search with a group tags that are more applicable than “Knowledge Management in K-12 Education”, even though this group may contain some of the documents that I am looking for.

As I mentioned before, there is no right or wrong way to tag, but from my perspective there are some ways to tag that work better than others.

thanks.
I think i can see what you mean, although i don’t see why you would need group tags for this at all.
you were using terms like “small database”, “many groups” etc. what is considered a small database/groups/documents?

which is the reason i created another thread asking if it is possible to abandon group tags altogether.

For the most part, I am not using group tags. However, I don’t dismiss them completely as other user may find great value in using them. I am only pointing out, based on my experience, how group tags may be used and invited others to share how they may be using them differently from what I observed. The only use I have for group tags is for some groups in my Mobile sync group that are, by design, temporary buckets. I’ll add a group tag from the sync group to add the documents to DEVONthink To Go for reading while I am away from my Mac.

There is no magic number where the classification results are optimized. DEVONthink just gets better at classification over time, as the database grows in size.

I’ve posted a link in that thread to reply that Bill posted some time back that may be helpful.

Great, thanks everyone. I’m still hazy about multiple different tags with the same name. If the name itself doesn’t delineate the relationship between for example, a tag and group, what then does?