FINALLY (almost...)


just discovered pb 8 via the newsletter - better tagging support is great news - can’t wait to see if the days of cmd-dragging replicates to folders are gone! :slight_smile:

good news #1:Tags exist. drilling into the group structure just became a bit easier, even if it is, well, a bit cumbersome still (why smart groups ?) -

wait, all of the german umlauts have rutsched to the down!

color: does it have to be grey (for a “Mac-Person”)?

and the tags are assigned according to…

good news #3tags.jpg :wink:

The gray tags represent tagged documents in groups (folders) that have not been excluded from tagging, while tags that you create will appear as blue.

got it, thanks !

It took me a while to discover how to enter these tags. These “Tag line” is a nice feature, although well hidden at present, I also like the new view very much.

At last, the info panel fits on my screen:

What a mess!

so, on a given rainy afternoon, I can tag along until I see all blue.

My issue, which may exist simply because I organize strangely is that I have lots of tags with duplicate names. My folder structure is this:

I have high level groups to organize by major sources: NGOs, Newspapers, Academic journals and books. I have similar folders within each: aid, civil war, etc. I was really hoping that in tag view I would be able to select the a tag and see all items tagged aid (a logical OR on high level each folder) but instead the tags are duplicated.

When I select more than one tag it shows a logical AND, which is even worse because selecting all of the aid tags shows me nothing (as no item is cross posted in all folders).

Anyone have thoughts on how I should be reorganizing or how tagging should change going forward?

Also, it is worth mentioned that I think this beta marks a huge improvement over the previous one. Kudos to DT.

If your group structure is


civil war


civil war

With tags you could have groups


And use tags “aid”, “civil war” on documents in those groups. In tags view, you can choose Newspapers AND aid, or NGOs AND aid. (You could create smart groups when you need a more persistent view.) In other words, what you have doing with subgroups you could do with tags.

That system might work. What surprises me is that the order of the tags matters (or at least it seems to), but there is nothing to alert the user to this.

For example tags in this order:
[Academic] [Aid] will put the item in the folder “Aid” which is inside the folder “Academic.” However when I look at tag view “Aid” will appear to exist as an independent entity, when it in fact only relates back to the nested “Aid” folder within the “Academic folder.”

I would much rather use commas for this nesting and have the tag appear as a single unit [Academic: Aid]. That, to me, conveys the proper meaning. These tags pretend to be flat, when in fact the order creates a hierarchy.

Does that make sense? Can someone from DT (or someone who has figured this out) explain how the (flat?) tagging system maps on to a hierarchical folder structure.

I’m glad to see that tags have arrived and that DevonThink 2.0 is moving somewhere to being close to near finished and out of beta after a year there. At this time I’m not so sure what I would use tags for exactly and there is at the moment a lack of input from the usual power users who often share very helpful advice. When people ask questions this if often the same collective of a dozen or so people who provides some kind of direct answer, in my own notes that brain trust would include Christian Grunenberg and the ineffable Bill DeVille of course, Annard, Greg Jones, Twicks, Rollo, Patrick Kroupa, SJK, Korm, Andreas Schmidt. Erich Boehnisch. There are many others who have posted information that I find very helpful of course, but those are the names at the top of my saved tips folder for using DevonThink. Also thanks to Mgrimes whose power user tips I’ve saved for DevonAgent reference and Houthakker for the amazing Devon to OmniFocus and back again scripts, I have not made use of any of this information yet since I use neither OmniFocus or DevonAgent but if I ever start your help seems to be fantastic!

I have to presume that some of you are having these conversations on some private beta section which is not accessible to the public where the initial betas are discussed and I fully realize that tags were only introduced today, but I see a distinct lack of input from most of the usual posters with the exception of Korm who posted this example.

Do the Devonian power users not use tags? Do most of you simply not care and tags are present to satisfy the general users of DevonThink while the power users yawn.

I hope the case is that people have not yet figured out how they will use tags and as they progress with working on their own databases would share their insights and tips with others, because right now I’m mostly scratching my head and wondering what good seeing rows of the same directory name scrolling off to the side which helps me find nothing at all is going to do for me. I also share the same concerns as rbiggs in the previous post. I understand tagging is undergoing major changes but the present version is slightly half-baked at best.

I am hardly a power user, just a “general user.” I am especially glad to hear that DT is going with OpenMeta tags. Lately I have switched to indexing certain files - e.g. articles imported via Papers, plus longer notes which DT’s AI engine would probably not be as efficient on. I am already tagging these on an ad hoc basis outside of DT with OpenMeta tagging software (e.g. Leap), and it will be nice to have the option of using those tags inside DT if I so wish.

I do hope the tags interface inside DT gets less clunky as things evolve. For example if the tags continue to be displayed vertically across the top of the pane that shows an individual file, then the height of the tag strip should be not just one row but two or three - as many rows stacked together as to hold all the tags in view.

I opened a new thread with a more significant title for this discussion on tags and re-posted some of my thoughts from the internal beta forum there:


This is what is going on. First, keep in mind that documents inside of tags are replicants - always. There appears to be an unintended consequence of the DT design decision that tags=groups. So, if you have a preexisting group (XYZ), and you add a tag (“XYZ”) to a document in some other group, the document will be replicated to the preexisting XYZ group. IF, however, the preexisting XYZ group has been flagged as Exclude from Tagging (see the Info panel), then a new group XYZ will be created in the Tags group. It gets strange if you later Un-Exclude the original XYZ group from tagging. The replicants in the Tag > XYZ group do NOT get re-replicated up to the other XYZ group. There are other even stranger variants of this scenario.

Clearly, tagging has added behaviors to DTPO such that the sequence in which you do something matters - this has not been the case before tagging was introduced. It is something that the designers need to carefully evaluate.

Tags look great and all of my tagged files (Tagit, openmeta) appear correctly in pb8. Will “See Also & Classify” include tags eventually?

After further testing of pb8 I have to add, that the tagging feature does not enhance my workflow at the moment. What I expected from tagging was a better way to organize and filter the content. In Zotero, for example, when I click on a folder, only those tags are displayed that are related to items in the folder:

This view provides a quick overview of the relevant content types. But even more useful is the following feature: Selecting a tag in Zotero not only filters the relevant database entries, it also filters the tag list itself, so that you see only those tags, which are also related to the selected tag. Thus, it is possible to “drill down” and focus on specific aspects of a given topic:
I could drill down further, and select another tag as well, e.g., Neurasthenie, in order to get articles that match both tags. But perhaps most important, I can see at a glance that some sociological articles in the “urban literature” folder deal with the more specialized topic of Neurasthenie. In pb8, I would not be able to see this, since the full list of tags is always visible.

This feature would be highly useful in DevonThink as well. At the moment, the new tag view displays every tag in the library, regardless which tag is selected. At present, the tag view looks too chaotic to be useful; even if I take the time to tag documents it would not help me to focus or to organize. This is unfortunate, since the tagging feature has the chance to improve DT’s ability to assist the user in finding relevant data. A filter would be much appreciated.

Still, pb8 is a huge leap forward - DT is clearly evolving to become even more professional now.

I believe that one of the greatest challenges to tagging in DT for existing users is that it can be difficult to use tagging in parallel with the established organizational/retrieval structure in databases that are already populated with thousands of documents. Also, until the entire feature set is in place in 2.0 to fully realize its potential, and limitations, I would be reluctant to think about fully transitioning to tags. I don’t mind sharing a couple of posts that I made in the beta forum-one which shows some of my frustrations with tagging and one which shows how excited I am about the potential of tagging. The two posts were made as replies in a conversation, so they may appear slightly out of context here.

Posted November 29

As I continue to test and work with the latest tagging implementation, I am impressed with the functionality and, at the same time, frustrated with its limitations. I’ll preface my comments with an acknowledgement that I realize that the DEVON group has always maintained that tags and groups would, for all practical purposes, function the same in DT 2.0. Having said that, I hope that tags will evolve beyond what is possible now.

On the positive side, I do like that I can tag in the tag bar and info pane-it is so much easier to ‘replicate’ documents using a tag. As mentioned by another user, the ability to tag/change tags on a group of documents would be a welcome addition.

On the negative side, once tags have been applied there appears to be little that can be done with them beyond what we have always had with groups, duplicates, and replicants. I really want to be able to perform searching and create smart groups based on tags, and I am not finding any way to easily do this. I was hoping to create hierarchal tag groups in the tag view, but that’s not possible. I was hoping to search on tags (only) in the search pane, but that’s not possible. Then I thought I could focus on a group in the three-pane view and switch to the tag view, limiting the tags to only those contained in the focused group, but that’s not possible. As tagging functions now, if I want to search for documents that have TagA AND TagB AND TagC, I have to scroll down the long list in the tag view, manually selecting each tag. I can find what I am looking for quicker by navigating groups in a standard view.

I’m curious how others are working with tags as currently implemented-do you all find similar results or are you selecting/searching on tags in a way that I have overlooked?

Posted December 10 in reply to the thread Johannes referenced above

It’s important to say up front that restructuring and tagging a legacy DT database that has been in use, and structured, and refined over the years is not very appealing. Having said that, I believe that one of the most significant things that tags brings to the table is that highly structured grouping with replicants, and to a lesser extent comments and keywords, become secondary. As example, consider this small sub-set of a database for someone in the field of knowledge and learning management. Without tags, I might have the following database structure:

Knowledge Management

KM Research * See naming note below
KM Case Studies
KM Theory & Application
KM Best Practices
KM Resources
KM Systems
KM Conference Proceedings

Learning Management

LM Research
LM Case Studies
LM Theory & Application
LM Best Practices
LM Resources
LM Systems
LM Conference Proceedings

With the above, when I get a new journal article by Kurtz concerning his research on a KM system to be presented at a conference next month, I will need to file the article to KM Research and replicate the article to the groups: KM Systems and KM Conference Proceeding. I read in the article that Kurtz has also done significant learning management work for the Alliance for Technology Access (ATA), but what do I do with that info? Do I create a new, top-level group for ATA or do I create an ATA group under Learning Management? Either way, I am stuck with classifying this document until I create a new ATA group somewhere. Once I create the group, wherever that may be, I go back to the original document, clip the relevant ATA passage, and create a new document in the ATA group. If I have created the ATA group at the top level, how do I link it to the Learning Management group? I don’t mix documents and sub-groups in the same group, so I need to decide to what Learning Management sub-group(s) the document should be replicated. Are we dizzy yet?

With a tagging strategy, my database group structure may look like this:

Knowledge Management
Learning Management

Case Studies
Theory & Application
Best Practices
Conference Proceedings

Now when I get a new article on KM systems from Kurtz, I file it in the the Knowledge Management group and tag it with Research, Systems, and Conference Proceedings. As the tags are suggested to me as I begin to type, data entry is faster than manually replicating to the groups. I clip the relevant ATA passage, and create a new document in the Learning Management group-I’m not mixing documents and sub-groups in this group, so this works. I also create a new tag for ATA on the fly-no need to think about where an ATA group should be structured in the database and no need to navigate through the database hierarchy to create the group.

It is also easier for me to tell where a document has been ‘replicated’ to via the use of tags as the user can set the option to have tags displayed below the navigation bar. There is no need to ‘Show Info’ and click on the Instances field to see where actual replicants are located. It is also easier to remove tags from a document than it is to navigate to groups to delete the replicants.

One thing thing that is still needed for tagging in DT is more robust searching capabilities. I would like to bring up the search window and search on tag criteria, and typing the tag names would auto-fill based on the existing tags, just as it works to assign existing tags to a document. Dragging and dropping tags into a search window or the search field in the toolbar would be welcome also.

  • As a note on group naming, if one uses tags at all, and uses group names as a tag, then it appears it will be necessary to somehow identify to what group the sub-groups belong. Otherwise you end up with multiple groups in the tag view that have the same name.

Yes! This is what I meant in another post. Something like this would be great in DT.


Greg_Jones, you described my problems pretty well in your second post. Did you get any useful responses to those issues?

I believe the response is that Dtech is listening to user suggestions and experiences. Dtech is one of, if not the most, responsive developers that I have ever seen, so I’m confident that tagging in its final form will work very well.