Even though it’s been pointed out here, in the suggested links, and elsewhere in places easily found – how to do all these things in DEVONthink, I finally realized that is not the point of this pointless thread.
I’m with jamesmicfish on some (but not all) of the points mentioned above, and I understand what he is trying to say, unlike some of the other commenters. Some of you seem to think that DEVONthink is perfect as it is, but it’s not from our perspective. You could listen to us and see an opportunity here to make the app usable for more users, or you can call this thread “pointless” (as korm did above) and lose users. DEVONthink has been working fairly well for me, but I agree that what jamesmicfish has said points toward some optional features that would improve the product for users like us.
While DEVONthink does use Apple’s file system to store Imported document files within the database, in the folder named Files.noindex, those document files are not organized into folders corresponding to the user’s organizational scheme for a database. There’s a good reason for that.
In the Finder a Mac user works with folders and files. Folders can be opened to view their contained files. Files can be opened and displayed under an application capable of viewing/editing them, or under Apple’s Quick Look. Folders and files may be duplicated or aliased.
In DEVONthink a database user works with groups and documents. A group can be opened to display its contained (located) documents, so is really metadata about organizational location. A document can be opened to view it within the database via viewing code for its file type built into OS X, or a Quick Look plugin or opened externally under another application, so is metadata pointing to a file. Groups and documents can be duplicated or replicated. Unlike an alias of a file, which points to the original file and remains as a useless artifact if the original is deleted, all replicants are pointers to the original group or document, so that if one of a pair is deleted, the ‘original’ remains, no matter which was deleted.
Replication makes it possible to organizationally locate a document within multiple groups, which can be very useful. The memory and srtorage space overhead for doing that is small, unlike duplication of the document into multiple locations. And if any replicated instance of the document is edited, all instances are edited.
Suppose I’ve replicated a document into 4 groups. If I export that database to the Finder, the replicants will be translated into duplicates, increasing storage space. If I were then to edit one of those duplicates, the others would not change, but would have to be individually edited or copied if I wanted them all to change.
I find replication very useful for a number of purposes. The inability of the Finder to honor replication is one of the reasons I don’t want my databases to be mirrored in the Finder via Indexing. I can do useful things in DEVONthink that would be lost be in translation to the Finder.
Bill, I am wondering whether you also use tags? Tags also make it “possible to organizationally locate a document within multiple groups” (where a tag is a group), both in DEVONthink and in the Mac file system.
For sure. That’s why I said that tighter integration with the file system in indexed folders would have to be an “optional” feature, not a replacement of existing functionality. There’s no point in dropping or replacing functionality that users depend upon.
I have absolutely no idea what anyone thinks about anything; so that kind of insight would be enviable. But DEVONthink is not perfect for sure. Pretty good; pretty flexible; pretty workable; takes effort; not perfect. What is?
We are in total agreement on that point! I think there is at least one good idea for improving DEVONthink in this thread. That makes this thread worthwhile and not pointless.
There have been questions or problems with indexing in the past. Many have been discussed (by me and others) over the years, and this includes how folder/group changes are handled.
I think where I disagree with jamesmicfish is that he has not taken the time to explore the rich treasure trove of knowledge in the forums, in the manuals, on the Internet, etc. Yes, DT could be better (everyone agrees). Yes, it is good to exchange ideas (everyone agrees). But, rather than trying to understand what DT is doing first, recognizing that there are reasons why things work the way it does, and then making suggestions for improvement based on DT’s vision for the app, he is trying to impose a radically different utopian vision onto it of some other app that he wants to see someone make, and I am not even remotely interested in seeing DT transformed into this thing.
This began as a question about importing vs. indexing, but has morphed into a bunch of suggestions for upending everything. I’d be much more interested in talking about more concrete ideas (related to the original post) and ways that we could improve certain aspects of the app – I’m not so interested in seeing a complete overhaul of it. I like how indexing works now, but I would also like to see an option (not a replacement for the current system) for having the DT database more closely mirror the indexed folders (for example, moving a folder or file in Finder that is moved in DT). In that, we are in agreement.
I don’t understand what this sentence means in this context, perhaps because I can’t parse what “things” means in this context. (Error: Variable “things” not defined!) I’m guessing it’s a statement of opinion that indexing of external folders in DEVONthink is perfectly fine as it is and there should not be an option for it to be any other way?