Request: Nesting Smart Groups

DEVONthink Feedback
1

I posted this a week or so ago in the troubleshooting section, and am now submitting this as a formal request: allow the nesting of smart groups.

Currently, the way the smart group function works, is completely inadequate. :confused: Details below:

Now that tags are here, I really want to push toward a flat file structure, with all groups, categories, etc. largely governed by tags. I’ll dump info into a small number of meta-level categories, and use tagging for the drill-down. At least, this is what I want to do. However, I can’t seem to be able to nest smart groups. Why not? I’d never tried it before in DTPO, but assumed the functionality would be there.

E.g., I can have a meta-level folder “School,” where I keep ALL school docs. I then want a smart group for, say “Contracts.” No problem. But then, what if I have a huge number of docs (notes, papers, presentations, etc.) in that group that I want to further define into Notes, Papers, Presentations smart groups? No dice, unless I want all of those smart groups to exist as another branch on the “School” tree, so to speak.

The point of a smart group is to “smartly” search all of one’s content based on certain parameters, and then aggregate it into one place. These smart groups, however, should allow for increasing degrees of precision. If I have an “Inventions” folder and want to smart group all my “Widgets,” but then further sub-classify “Widgets” into “Widgets-with-wings” and further “Widgets-with-wings-with-wheels” and. . . etc. This process should be allowed to continue ad infinitum.

However, in the current setup, my “Widgets,” Widgets-with-wings," and “Widgets-with-wings-with-wheels” smart groups are all at the same hierarchical level within the “Inventions” folder. This is illogical and counterintuitive, as all smart groups after “Widgets” (in this example) are a sub-classification of that “Widget” class. So, the folder/search hierarchy should be thus:

Inventions[regular folder]>Widgets[smart]>Widgets-with-wings[smart]>Widgets-with-wings-with-wheels[smart]

Each of those should be nested within the preceding item in the hierarchy.

This is the beauty of smart groups. They can be as precise or as imprecise as one wishes, and they are not permanent–they can be wiped away without a second thought, when they cease to be useful (unlike a traditional folder hierarchy, in which everything would have to be physically moved).

So, please allow the nesting of smart groups.

2

celsee has my vote, fwiw. I like iTunes v9-style smart groups - DT’s smart groups are so last year. :frowning:

3

I would rate this as very important along with improving the current implementation of tags.

For example not even iTunes 9 style smart groups which should be the standard but iTunes 6 or 7 smart groups at the very least.

I can’t seem to select a date range, or for instance: date is before year, or date is after year. If this is somehow possible using the boolean operations it is not in any way obvious to me how to accomplish it. I would like to do something relatively simple like search for all documents added during a particular time, my only seeming options right now are the pre-defined searches of: today, yesterday, last week, last month, this year, last year, which is not nearly detailed enough to be of much use.

How do I search for: date range is between 2003 and 2004 for example? Or date of file is before 2003. These are simple, non complex searches which I cannot do. If I can do this somehow it is not at all obvious to me how.

The help manual states the smart groups are still being developed so I hope there are plans to at least bring in iTunes from 2004 functions, but of course iTunes 9 would be the standard for truly flexible and powerful smart groups as celsee and korm described.

4

Something like Journler’s Smart Families?

5

I’m not familiar too much with the Journlr interface, it looks a little strange, it’s a piece of discontinued software no?

If it does the same function as iTunes 9’s any or all of the following

and … dot dot dot

next series of any or all of the following (nested smart groups)

Then yes.

If it’s doing something completely different that I don’t understand, then not really :open_mouth:

I agree with the previous posts, iTunes 2005 or something like that would be ok, being able to select a date range. iTunes 2009 would be the gold standard :smiley:

Project name is, tags are, date range is between March 25 and May 15, 2008. Simple stuff that’s impossible right now.

6

If you regard regular folders as if they were tags, and that presence within a folder is what marks something as being “tagged”, then with replicants you could implement exactly what you’re asking for, could you not?

John

7

Yes, it could emulate the behavior, but I disagree that it is a good solution. I’m looking for a flat file structure, not an almost flat file structure.

To me, it makes no sense that they don’t nest. They should behave, as sjk indicated, much like the smart families in Journler.

With smart groups and the ability to tag, I don’t want a workaround for what could be normal behavior. On top of that, I still have to create actual folders “pretending” to be tags to build the hierarchy, not to mention the added step to replicate after creating the folder. . . I could’ve killed two birds with one stone whilst defining a smart group.

FWIW, I think the new replication structure is way better than the ridiculous color scheme that used to exist, but it still is not an ideal substitute for actually nesting smart folders.

So, to that end, I hope smart groups are an evolving species for Devontech, in that they don’t continue to behave as they currently do.

8

And I’d also like to add what is perhaps the biggest issue (impediment) of your proposed workaround:

Tagging is not replication. I tag my files as I work, over a period of weeks, months, years. When I want to make a multi-tiered smart group, those tags are already there. . . all I need to do is define my criteria for the group.

However, if I work on this level with replicants, I need to go through the process of re-replicating the file into whatever group I want it in–every single time. If one begins to repeat this process by a magnitude of 10, 100, 1000, one can see that it is tedious, and quite the opposite of efficient.

9

I do agree that replication can be quite manual, compare to some of the other tagging techniques around (I love browsing tags with Leap).

Could you not also emulate a “flat” structure by having a smart group that contains all records?

John

10

Agreed 100%. It isn’t can DevonThink do this right now, sure it can. It is however tedious and aggravating, exactly as celsee described.

11

I don’t want to emulate such a simple thing. I should be able to do it.

But, can you expand on what you mean? I’m thinking about a smart group with all records. Then what? Add folders to a smart group and replicate from there? I think that’s what you mean, but I’d like to know if that’s incorrect.

At any rate, again, this is nowhere near ideal. Programs far less complex than DTPO have this capability. I’m certain that the developers that can build an AI system like DTPO’s can code in the nesting of smart groups :smiley:

Incidentally, where are all the DT personnel in this thread? Bill? Christian? Eric? Anyone??? I’d be elated to hear something from the DT team on this.

12

@ celsee: Having just introduced Tags, the developers are monitoring user reactions and suggestions. Which suggestions gain support by other users, which don’t? Among changes that could be accomplished, some changes are possible, others may not be, or would contradict recommendations by other users.

And that’s true of smart groups as well, especially those that intersect with tagging issues.

13

I can’t imagine any user protesting the nesting of smart groups (rhyme inadvertent).

14

If that is the case, my vote is for them to check out Leap. It does implied tags (based on directory hierarchy) + explicit tags, just like DTP does, except the interface for finding tagged items is brilliant. You have two columns: one with your tags in it, and one with your group hierarchy. If you select a group, the tags column populates with all the tags extant for items in that group and its children. If you select a tag, numbers appear next to every visible group in the groups column to show which groups contain items having that tags. Multiple selection acts as a union operator, and of course smart groups involving tags are possible.

My guess is that since DTP maintains its own metadata database, the user interface could be even better than Leap, which gets really slow when tens of thousands of files are involved in these operations.

John

15

As DT developers stated that the tag view is still incomplete and will change ‘substantially’ (quoting from manual) I think it could be very practical to hear some news from them. I mean: I think everybody will agree that changes in tags view are needed so it could be useful to know (even in a very general way) what are developers plans with it. I agree that the Leap-way can be a good sample and I also think that a more robust inspector is a priority: as said by a number of users tag filtering is a must. I basically use tags to assign ‘qualitative’ keywords to quotes and notes for my research and so if I select the tag “Goethe” I want to be able then to skim results to get to, let’s say, quotes that talk about (tag)female characters and (tag)love (quite a stupid example, I admit!). I know that this can already be done by smartgrouping with tag1,tag2,tag3 but that’s, obviously possible only if you already know what are you looking for. If you’re just brainstorming the possibility to skim data from a macro-theme and see what it associated to this macro-theme is a must, I think.
So: anybody from DT can define a little more what “change substantially” mean?
Anyhow good job, guys
go ahead :stuck_out_tongue:

16

That’s something I also wanted to ask for. Perhaps the developer’s could give us some feedback on our feedback and tell us what their plans are. What sounds useful for them and what doesn’t. Perhaps (if this is not asking too much!) people with a special interest in tagging could be included in non-public beta releases which bring further development to the tags view.

I think tag-filtering according to selected tags is a much ask for feature. It makes all the difference in the usefulness of tagging!

Birgitt

17

Here’s another vote for nested smart groups in an iTunes or some such style. Essentially I would like to be able to mix “all” and “any” arguments in smart groups.

Thanks

Nick

18

Hover the mouse over the ‘+’ when creating a Smart Group and notice the Option-Click the + to create compound predicates. tooltip, which is the action to construct the result you’re essentially interested in. :slight_smile:

19

Aha, thanks! :wink: