Smart Group Features

Hi all…

I’m currently trying out Devonthink and from what I can work out I think it is missing out on a couple of simple but to me key features regarding the way it handles “Smart Groups”.

I currently use Journler which handles “smart folders” excellently (I love journler but am forced to look for a new solution since it now has major stability issues running on OS 10.6 and the developer has stopped working on the project). I use a system where I file everything in the same folder - apply tags to each item - and use smart folders to sort everything out. This allows files to automatically appear in multiple folders if they happen to be relevant to more than one thing (e.g a file tagged with “webdesign”, “inspiration”, “projectname” can appear in a smart folder set up to display all inspiration for webdesign and a smart folder to display all inspiration for a particular project.) I have read about replicants and don’t think they solve the problem for me. I just can’t see the point of going to the trouble in tagging things if I then also have to go to the extra effort of still finding somewhere suitable in a folder hierarchy to put them.

I find to run this system however that I need slightly more flexibility than Devonthink can offer when it comes to setting up rules for smart groups. Most importantly there is no option for “does not match”. I often need to set up a smart group that not only contains a certain tag but also does not contain another. (To use the previous example, I might want a smart group that displays files tagged with “projectname” but not “webdesign” to show all files relevant to a particular project that are not to do with webdesign.)

I would also miss what Journler calls “smart famalies” whereby you can nest smart folders within each other with the result that child smart folders inherit the rules of its parents. This feature adds a great flexibility to organise smart folders in a creative way and adds a hierarchy structure without which using a purely smart folders driven system can get really messy.

I’ve had a quick look on your forums and I have found several posts with people asking for similar features. I also wouldn’t have thought they would be too difficult to implement (especially the “does not match” rule)? So my question is would either of these things be in the pipeline for an update in the not too distant future?

Been there, e.g.:

here
here

:slight_smile:

Although there’s not currently a direct “is not” filter for a tag, there’s a way to do that.

  1. Using the full Search window, perform a query with the objective of finding documents that meet some set of criteria (but not referring to the tag).

  2. When the search is complete, select all of the listed results and replicate them to a new group freated to hold them.

  3. Run a search in that new group for the specified tag.

  4. Move the items that have that tag to the Trash.

Result: You have have a group containing replicants of all the items that DO NOT have the specified tag, but meet any other criteria you have set (including, if you wish, a list of all the other documents in the database).

This isn’t a smart group, but reflects the current content of the database.

Thanks Bill for the work around, but it seems to me very messy and still won’t leave me with a folder that will be kept updated with the latest additions to the database.

sjk - I hadn’t seen the “Nesting Smart Groups” topic - Its an interesting read and addresses exactly the issues that were concerning me - Basically that the current way DevonThink handles tags and smart groups is quite basic and is especially inadequate for a system which relies on them to organise itself.

The topic is ended however with reasonable questions regarding in what direction DT developers are planning to take tag and smart group development and when it might happen. The questions were asked way back in December so I wonder if anyone from the team is any closer to getting an answer to this? I think for myself and the system I run I can not look at investing in what otherwise seems to be an excellent product until it is a little clearer as to what is happening.

No surprise since you’re new here (hi), though you might have spotted it with a search for “journler” since that’s how I dug it up… with an obviously advantage of having previously seen it. :slight_smile:

That’s why it seemed worthwhile to reference here.

… without continuing to a satisfying closure. Glad you’ve brought it up again.