Forum Posts Feed
Any Known Alternatives to DEVONthink for Linux or Windows?
< Next Topic | Back to topic list | Previous Topic >
Posted by Alexander Deliyannis
Jul 13, 2018 at 08:21 PM
Paul Korm wrote:
>I’ve all but abandoned using it after
>realizing the file system and a few useful low-cost utilities gets more
>done than fiddling with databases.)
Can you mention some of these utilities?
Posted by Paul Korm
Jul 14, 2018 at 10:40 AM
Finder searches are far more adaptable and configurable than any DEVONthink search
Tembo for quick on-the-fly searching
Ammonite for tag searching
Apple Notes or Curiota for on-the-fly note taking
Forklift 3—lots of nice features for folder and document management
For me, the above replace 99% of what DEVONthink does. I don’t need a database to store documents. I don’t need fancy sync—iCloud is adequate. The See Also & Classify feature is interesting but I rarely use it. The amount of time spent on managing and curating folders in the file system is no different than the same task in a database. And every macOS app is integrated with the file system by default, but none of them are integrated with DEVONthink—if you use DEVONthink to store documents you are 95% locked into using DEVONthink to launch them for editing.
Don’t get me wrong—it is a nice product, well engineered and reliable. But for me it is also overhead with little payback.
Alexander Deliyannis wrote:
>Can you mention some of these utilities?
Posted by nathanb
Jul 14, 2018 at 01:56 PM
Paul Korm wrote:
Finder searches are far more adaptable and configurable than any
>DEVONthink search
>Tembo for quick on-the-fly searching
>Ammonite for tag searching
>Apple Notes or Curiota for on-the-fly note taking
>Forklift 3—lots of nice features for folder and document management
>
>For me, the above replace 99% of what DEVONthink does. I don’t need a
>database to store documents. I don’t need fancy sync—iCloud is
>adequate. The See Also & Classify feature is interesting but I rarely
>use it. The amount of time spent on managing and curating folders in
>the file system is no different than the same task in a database. And
>every macOS app is integrated with the file system by default, but none
>of them are integrated with DEVONthink—if you use DEVONthink to store
>documents you are 95% locked into using DEVONthink to launch them for
>editing.
>
>Don’t get me wrong—it is a nice product, well engineered and
>reliable. But for me it is also overhead with little payback.
>
>Alexander Deliyannis wrote:
>>Can you mention some of these utilities?
I’m on Windows so don’t know the specifics here, but am assuming that storing files in the file system vs a database are roughly the same.
I’ve long wrestled with this issue about where the metadata for a file should live. I agree that far and away the best place to store document metadata (tags) SHOULD BE in the file itself so the user isn’t depending on a separate index (database or not) for curation. Lots of file types have their own metadata standards built in (incompatible with each other, of course). Like adobe pdf tags, exif info in photos, and MS Office Docs even has their own group of tags which you can get file explorer to show….
...ah that specifically brings me back a couple years when I was trying to migrate a bunch of my company files to Sharepoint. I was surprised to find that ALL that embedded metadata for Office files beyond the filename was completely ignored upon import…. Sharepoint even ignored the basic created/modified dates. I was quite perplexed that Microsoft regarded what seems to be a fairly good metadata scheme on it’s own documents as not worth bothering to preserve. I mean, if MS is going to ignore this info upon migrating to a ‘more modern’ platform, then who IS using these tags? Que the MS jokes about abandoning a good thing for stupid corporate reasons (RIP OneNote) but this particular example is typical of the whole industry. Picasa being another good example. I use AutoCAD too , a horrific Frankenstein of 3 decades of different embedded metadata schemes that inexplicably still ‘coexist’ within the most recent version.
Examples like that over the years have made me cynical about relying on file-based proprietary metadata for ANYTHING large or long-term. Is this situation generally more stable in Mac-world?
Besides the trouble with changing metadata standards and maintaining a mixed file system of different standards, isn’t the other main benefit to use a database is to browse the scope of your metadata (push) instead of just blind search (pull)?
Take tagging pdfs inside the pdf (using the Adobe standard). If I copy a directory of those pdfs to a different pc, is there any software that scans those to show how many of each tag exist or do you need to explicitly (get lucky) search for them? Maybe I’m just missing a whole category of software. I’m not an academic but I’ve always had this impression that a solid granular metadata situation exists in the Adobe world that I’ve never been able to leverage because in the business and engineering world not enough people care about or even recognize document metadata as being useful.
The thing is, when I curate my own collection, I want to create a dashboard to it beyond ‘powerful search’. I want to be able to see the scope of my category and tagging scheme to help guide how I label the newest item but also to remind me what my whole library looks like and about the connections I’ve made in the past. My recall memory is garbage but my recognition memory is pretty good. So I need my collection to remind me about stuff I filed for a good reason a year ago that I’ve completely forgotten about today. Let’s say I ran across a great white paper “What Entomology can Teach Us About Information Theory” a year ago. If I file it in folders, the best I can do (i think) is to tag it with a bunch of keywords so it’ll be more likely to pop up in search. If I file/index it in a database then I can associate it with various knowledge graphs/maps and have it show up in a few different branches of tag/category trees. For me to rediscover it the first way, I’d have to one day actually search my own stuff for ‘ants and knowledge’ or something like that…which will NEVER happen because my recall memory is garbage. I’ll remember that article IF I run across it again and then the old idea connections it spawned then would spring up again with hopefully new insight today. But the only way…well, the most likely way…I’ll run across it again will not be upon search discovery, it’ll be through navigating a curated knowledge graph/tree.
Am I way off that it’s not possible to have the freedom of just files in folders AND a useful knowledge graph of those files? If I’m wrong, please point me in the right direction!
Posted by Paul Korm
Jul 14, 2018 at 03:30 PM
As with anything to do with “metadata”—it depends. Including the definition of “metadata”. The way Windows or macOS handles tags varies, and the way third party software on those platforms recognizes tags varies by developer. Tags are not the only form of metadata, either, just the one that most people are familiar with and like to chat about. In theory, a database can support the metadata managed by the file system and augment it with custom metadata, either defined by the developer or the end user or both. In practice, on macOS, that doesn’t happen much. DEVONthink has no user-defined metadata, and a few developer-defined attributes. macOS and Windows support additional attributes in the file system but these are not easily accessed by the end user.
nathanb wrote:
>I’m on Windows so don’t know the specifics here, but am assuming that
>storing files in the file system vs a database are roughly the same.
Posted by Chris Murtland
Jul 14, 2018 at 07:12 PM
The ability to apply metadata across diverse files and information types has always been the draw for me for database approaches like Ultra Recall. In UR, I can set up shared fields that are used on notes, PDF files, email messages, etc., and then use those fields in searches (and those saved searches are items like any other, with metadata also applied to them).
That said, the approach of something like Notebooks, which is entirely dependent on the filesystem, is also highly appealing to me. Not being locked in and being able to operate on the data using various other tools seems more sensible in a lot of ways. Unfortunately, tools based on the file system rarely provide any real metadata capabilities (at least on Windows), and therefore prevent much in the way of slicing and dicing.
I’ve never really resolved this tension to my satisfaction; wish I could have both. UR slightly wins out though, in that it can synchronize to external folders, so there is a possibility of getting the best of both worlds to some degree.
However, there is also a cost to maintaining fields/attributes/metadata on a lot of items. It pays off sometimes, but the file system approach seems to have less overhead.