I read in another discussion that the “next version” of DevonThink (I’m using Pro) will be visible to Spotlight searches. When will this be out? For me right now, the invisibility of my DTP database to Spotlight is one of my major reservations about the program. When this is implemented, will Spotlight find (and open) specific files within the database, rather than just telling one that a database contains the search terms?
On a separate note, could it be possible to make DT indexing work with aliases?
The next release (1.1.2) will be a maintenance release (fixing minor issues, improving the performance etc.). But v2 (supporting and using Spotlight) won’t be available before the end of autumn or winter.
While we are at it: will we just have the database exposed to spotlight or will we have spotlight search queries forwarded to DT which then spits out some search hits?
Why would you like to use Spotlight to query the database, since its queries are less sophisticated than DT’s? The right tool for the job so to say.
Most users just want to be able to have the file data accessible through Spotlight. Again, the purpose of this is also debatable but understandable since I guess they would feel less “locked in”.
Your comments are appreciated because the whole Spotlight integration is still a work in progress and feedback can help us get an idea why people want to use Spotlight with DT.
Well, I don’t even know if this is technically possible, but it would provide a nice “bridge” between spotlight and DT. So I would just start a spotlight query, part of the results will come from DT, selecting one of those will send me directly to DT’s search function/results.
It might be neat if there were additional options… like: I search spotlight for a file that happens to be in DT, and the DT results will have a group of “see also’s”…
If I understand Spotlight correctly it doesnt’t really do much smart searching - rather just some metadata and extracts from pdfs; whereas a DT search would give you much more than that… and integrating both would be nice…
No, this will not be necessary, not even come close to being that – it would just be really cool and make the DT database much more transparent and part of the whole system. You know, the kind of mac coolness and geekery
I’d say that depends on the type of query. Simple example: I can easily search for all PDF documents using Spotlight but haven’t found a way to accurately do that within DT databases (though it may be possible with scripting).
In general, I think DT has better content searching and Spotlight has superior metadata searching. Lack of the latter is frustrating for me in DTP when I want to locate items in databases by their document type, that have a script attached to them, containing wiki links, with replicates, etc. Again, some of that could be done through DTP scripting but that’s more challenging to me than writing raw Spotlight queries for mdfind.
Most people will agree that the search UIs for DTP and Spotlight could both benefit from improvements that would make it easier to take more advantage of their respective underlying technologies. It’s challenging to find any UI for sophisticated search queries that I’m fully efficient and comfortable with. I hope to get closer to that usability goal even if it’s never completely satisfied.
MoRU’s not too bad (feature-wise anyway) and I’d have been tempted to buy a low-priced ($10) license if its UI weren’t so intolerably sluggish on my iMac G5. Haven’t tried HoudahSpot, which costs $5 more than MoRU and I’m mostly broke right now. Or Meta, for another $5 bump. Or Desk Lamp, which clocks in at $25. I don’t remember others offhand, except for some specialized utils like SpotMeta. I occasionally tinker with the free NotLight but its UI is far from friendly. With 10.5 on the horizon I’ll wait to see what Spotlight-related changes it has.
The main trouble I have with Spotlight search UIs (at least) is they require awkward combinations of keyboard and mouse usage to do anything reasonably powerful. That may be tolerable for infrequently configuring Smart Folders and such but gets old fast with frequent on-the-fly searches. And it’s hard making quick adjustments to various query/filter criteria to help focus on specific desired results.
Those issues affect my DEVONthink searching in varying degrees, too. Selecting some desired combination of radio buttons and menu options is a cumbersome pre-search ritual, made worse by having only one saved state for them.
Surely there’s potential for UIs enabling more streamlined construction and modification of moderately complex search criteria that even relatively unsophisticated users motivated to learn could comprehend, especially if inspired by tangible results as examples. Phew. Bridging the gap between search possibility and usability.
As a little teaser: I just began a new Cocoa app called MetaFinder which does more than just search – it lets you browse your hard-constructed metadata.
For example, I now keep all of my downloaded software bits (about 55 Gb at this point) in one gigantic directory, without any structure. I only use directories to represent lifecycle relationship, kind of like “packages”: in that if I delete the directory, I want everything it contains to be deleted at the same time.
I use SpotMeta to assign custom metadata tags, representing the general category, the target OS, the language, the coding framework, the game genre, etc. Within these tags, I use / delimited path strings, like “Applications/Utilities”.
Now, what MetaFinder lets me do (and it works well at this so far, although only in read-only mode) is to select one or more metadata tags, and it then builds a virtual directory hirearchy based on the contents of the chosen tags. If I want I can also specify a regular Spotlight query, and it will constrain the results to only those entities matching the query. But in its default mode, it will show me all of my applications, for example, structured within a hierarchy represented by the path strings I typed into my Category metadata field.
It’s like having the same file in multiple directory hierarchies, whose structure I determine at query time.
I’d love to queue up some alpha testers… (hint hint). It works in read-only mode right now. The next step is to support drag-and-drop, so that if you drag-and-drop a file into a Metadata virtual directory, it will automatically apply the appropriate metadata so that the same item will appear within the same virtual directory the next time you make a query.
When I search for something on my computer, I use Spotlight, but then I also have to go into DevonThink and repeat the search… this is annoying! Especially for a tool (DevonThink) whose main feature is searching across your info.
I’ve considered using Index instead of Import, because then Spotlight would see the documents; but that won’t work for me because I really need the ability to edit things in DevonThink, and move quickly between documents in DevonThink.
Perhaps I should try doing Export on my entire DevonThink database? Then Spotlight would see those files. This is not ideal because: I would need to do this daily to ensure the export files match the DTDB; I must be careful to not edit the exported files; Spotlight would bring me to the exported file, not to the file inside my DTDB.
Well, for now I mostly remember what I have inside my DTDB, and so I usually know when to search in DT vs. when to use Spotlight. But that’s relying on my very fallible non-computer brain!
I am very attracted to the whole idea of using metadata to organize and search files; I haven’t started on this yet, but I hope to soon; however the DTDB world is going to be cut off from any Spotlight related metadata system (like SpotMeta, Spotlaser, etc.).
I can see that this may be a hard problem for Devon-tech to get Spotlight to regard pages in the DTDB as separately searchable and openable files. But I hope that Devon-tech maybe gets a better understanding from this post of what one user wants anyway.
I wonder if FoxTrot <http://www.foxtrot.ch/foxtrot/> would fit your needs. I’m still using OS 10.3.9 so I’m not up to speed on the pros and cons of Spotlight. But I find FoxTrot pretty useful when I want to do a search that will include files in my folders, DTPro, and–as a bonus–my e-mail (I use PowerMail). It assigns a rating to the search results, you can see a preview of a particular file before you open it, and the search terms seem to be more flexible than DTPro allows at this time.
Thanks for the suggestion about FoxTrot. How does the integration with DTPro work? I didn’t see specific mention that FoxTrot works with DTPro. But I did see something about a “FoxTrot crawler/importer to get extend FoxTrot to support additional document types”… Perhaps someone has done this for DTPro? Does FoxTrot work with DTPro out of the box?
Well, I may have been a little hasty in suggesting FT as an all-around solution. You can specify the folders it will index and, while I have my main DTPro database in Documents, it looks like FT only finds PDFs that are related to my search–none of the rtf or web archives I know are there. I prefer to import my docs, perhaps if Index were used the results would be different. And I gather that the DTPro v2 database format will change; so that may allow a more thorough search.
I wonder if it would be worth while to ask CTM Development if there’s any chance to get FT to look inside the DTPro database.
Should we expect to see a Spotlight plugin available any time soon. As I’ve begun to use more meta data with my files, I find that I use Devonthink Pro less. I don’t want to do two searches to find my relevant information. Please bring spotlight integration to DTP.
How does the integration with DTPro work? I didn’t see specific mention that FoxTrot works with DTPro. But I did see something about a “FoxTrot crawler/importer to get extend FoxTrot to support additional document types”… Perhaps someone has done this for DTPro? Does FoxTrot work with DTPro out of the box?
Foxtrot comes closest to (the still more robust but obsolete) Ultrafind of old. It would be ideal to add its functionality to DT.