Are there any disadvantages to indexing a folder of documents (web archives in my case) vs. importing them in DTP?
I would like to have the files easily accessible in Finder and by other apps, so indexing makes more sense, but I am wondering if there is some gotcha about indexing rather than importing?
The only snag I can see is that I have to synchronise the folder to get updates into DTP, but that doesn’t seem too burdensome (or does it become slow when there are lots of documents involved?).
Also… is there a way of having DTP synchronise automatically (either periodically or, ideally, whenever I sync to DTTG on my iPad)?
Rather than point you at one or the other posting, I’d suggest you use Advanced Search (up there in the right hand corner of this page) and look for anything that Bill DeVille wrote on this topic as a starting point. If your research gets you nowhere, then perhaps at least your next question in this thread will be informed by the prior art on the topic.
I had already drafted the info below before korm posted the helpful link to Bill’s notes, but I’ll post this anyway.
The disadvantages are that a) all of your documents are not contained within the database for backing up, moving to another Mac, etc., and b) if you move the documents/folders, or rename the folders, in the Finder, then the index link will be broken. The downside of ‘b’ can be minimized somewhat by reducing the number of folders that are indexed, and subgroup folders under the indexed folders. In other words, a Finder folder structure of:
Indexed to DEVONthink folder
Folder A
Folder B
Folder C
is preferable to:
Indexed Folder A
Indexed Folder B
Indexed Folder C
Not slow at all. DEVONthink also includes a Synchronize script on the disk image that can be attached to an indexed folder, and the contents will be synced every time that folder is selected in DEVONthink. In the above example, you would attach the script to ‘Indexed to DEVONthink folder’. Scripts are attached to indexed DEVONthink groups from the info pane.
I suspect that this would be scriptable, but as mentioned above, the groups will sync when activated if the Synchronize script is attached.
The use case @markauk describes needs clarification - but if it means that a DTTG sync first triggers a desktop sync event between DTP and an indexed folder, before the DTTG <-> desktop sync event, you’d need a script at the system level that listens to Bonjour, looks for a sync request from DTTG, hijacks it, causes DTP to sync its folder(s), then lets the Bonjour request proceed. Even if that’s feasible, it’s probably too daunting a task to try. Sync with DTTG (or anything else) cannot be initiated from the desktop, in the current iOS architecture.
re triggering a DT sync when syncing from iPad, I guess I was hoping there was a way that DTP could, when it got a DTTG sync request, first do a folder sync. I definitely don’t want to go to hacking something which listens to Bonjour etc!
re questions being informed by prior art… I did spend quite some time looking through the online documentation and searching the forum, but didn’t find anything which helped. One of the problems with a very active forum, a complex piece of software and someone who is new to the SW and its terminology.
Glad to help out. There is one important consideration of an indexed database that I forgot to mention. If you create documents in DT or capture data to DT via the various capture tools, the data does not sync back to a Finder folder. The non-indexed files in DT can be placed in a group that is indexed, but the files remain in the DT database.
Good point, thanks. For what I have in mind this won’t be an issue, and I suppose if I did ever want DTP created documents back in a finder folder I could just export them…
After reading many of your other replies in the forum I was wondering if this has perhaps changed since 2010 with the current version?
This would be my dream use of DevonThink. Maintaining the original Finder structure because of its large 40+gb size, yet be able to index it and use all of the great DTP features such as date new from template, tags, search, etc. Also having new data/notes/files, etc. be copied to the original Finder location and not maintained in the DTP local database would be great. This would also eliminate the duplication of all of this data to the Macs in our small office. Currently is now setup up with local sync store on our Mac Mini server. Anyway to do this with scripting or am I still out of luck?
The one thing that has changed since 2010 is that a synchronize script is not needed to add new documents as DEVONthink now automatically indexes the entire contents of indexed (Finder) folders to the database. The user will still need to move documents created in the database out to the indexed folders, but I do not find this to be a major shortcoming. I still index nearly 100% of all the documents in all of my databases, and it works very well for me. However, I do not do this to mirror the same document structure in the Finder. To attempt to do so will eventually frustrate you if you replicate documents in your database, and replicants are extremely useful.
No it won’t. If you are still using Sync it will Sync the indexed data as well. Indexing will not make it accessible to other Syncing Users consuming less resources than the curent method.
Here’s what I would try: Don’t Sync the 40GB database. Host it on one machine with Web Sharing so people can access and search it if need be but not have to maintain a local copy that is Synced. (Note: This is based on me not knowing your full environment and needs.)
I would probably just Index it and not sync, possibly would sync the laptops though
Haven’t played with Web Sharing. Can the clients also add/delete files? I will have to investigate the tradeoffs of the web client vs. the full DevonThink client.
If the database is not shared as Read Only they can add and delete files. Yeah, check it out. It’s definitely a lite version compared to the app but it just might do what you need.
