2

Possible Duplicate:
Where to find a tag based file manager?

Anything similar to this available? I have run into limitations to organizing requests into various nested folders. I am interested in the ability to organize files in various ways and keep track of them over time. In some sense I would like something like a project management system but actually built into the filesystem/browser so I can save files from various tools and easily organize by contact, organization, topic, createddate, file groups, etc without having manually upload files to a tool. Browsing files later I'd like to be able to pull everything back up and be able to view in various ways for example organize files by day and see what was done on a certain span of time, or see all files for a contact, etc.

Community
  • 1
Dan Roberts
  • 281
  • 2
    This is exactly what WinFS was supposed to be, but the project seems to have been put on the back burner by Microsoft. Specifically, you're looking for a Relational File System, as opposed to a Hierarchical File System (what most everything is currently) – Darth Android Jul 19 '12 at 14:06
  • @afrazier: not really a duplicate since the question refers to a lot more than just tagging. – Julian Knight Jul 19 '12 at 17:28
  • Yup; this has been a massive thorn in my side for years. Do you put a photo of a cat playing the piano in the Cats folder or the Funny folder? Like Darth said, WinFS was supposed to be (one of the first) database filesystems which would have at the very least helped to resolve this sort of conundrum, and it was supposed to be part of Vista, but it has been all but cancelled, leaving us still looking for a way to sort and organize files. :-( – Synetech Jul 19 '12 at 17:38
  • @JulianKnight: I'm not seeing it. Everything in the question refers to tagging files with bits of metadata and then being able to filter/browse the filesystem based on the metadata. Inferring that the metadata should be key-value pairs instead of simple values doesn't significantly change the nature of the request. – afrazier Jul 19 '12 at 18:00
  • @afrazier: I guess that to me a tag is a form of meta-data not the other way around. A tag is a way of adding some meaning like "cat", "blue", "sunset". Meta-data is wider than that and includes dates, geolocations, etc. Semantics perhaps but this seems to be the accepted meaning. – Julian Knight Jul 19 '12 at 19:47
  • @Synetech: I know what you mean. The thing is, it's not even that hard! Picasa does it for photos, there are standards for meta-data in NTFS stored files and have been for years. IFF file types on the Amiga had standards for meta-data! LDAP does it. We even have the "big data" stores to handle it (MongoDB for example). Very frustrating. – Julian Knight Jul 19 '12 at 19:50
  • @afrazier: So to me tags are not really key/value pairs but a list: {tags: ['cat', 'blue', 'sunset']} (to put it in JavaScript terms). – Julian Knight Jul 19 '12 at 19:53
  • @JulianKnight: Your definition is more correct, but the most common requests in this area (even explicitly mentioned in this question) revolve around tagging and browsing by tags. For more generalized metadata searching, Windows Search built into Vista or newer and Explorer's Saved Searches can handle this very gracefully. The hard part is associating the metadata you want with the files. – afrazier Jul 19 '12 at 20:47
  • @Synetech: I make gratuitous use of hard links and (to a lesser extent) symlinks for file organization. No need to choose if a file belongs in a particular category or another, just use both! :-) – afrazier Jul 19 '12 at 20:51
  • @afrazier, Where do you put the actual file? What happens when you want to delete a file; you manually delete the file and each link? What do you do with your FAT drives? Hard-links are just a work-around and not a substitute for a proper implementation. – Synetech Jul 19 '12 at 20:54
  • @Synetech: The files just get hardlinked to wherever I need them. Deletion can be a bear, but my link counts generally stay low enough for it not to be an issue. And I generally don't need to ship and entire tree around enough to worry about propagation issues (or FAT drives, which I've long since given up on for anything but flash media for devices that can't handle anything else). Links aren't a perfect solution (in particular, management doesn't scale to high link densities without lots of help), but they can be an acceptable one for certain collections. – afrazier Jul 19 '12 at 21:00
  • > The files just get hardlinked to wherever I need them.   @afrazier, yes, you can put links wherever you want, but you still have to store the actual file somewhere specific. Do you keep one single, massive "dump directory" with all the files then link them to where you want, or do you (somehow) select somewhere to put the file (in which case you're back at square one: Cats or Funny)? For sorting and organization purposes, hard-links are effectively no different than simple shortcuts. – Synetech Jul 19 '12 at 21:10
  • "Do you put a photo of a cat playing the piano in the Cats folder or the Funny folder" - no, I put it in the "2012-06-15 Playing With Cat" folder (assuming I took a load of pics - just "2012-06" folder otherwise) with IPTC tags of "cat, piano, funny" and an IPTC. description. That way, I can search with a variety of tools including Picasa. Often I wont need to search because I remember roughly when I took it. All photo files are named as -.. But this is a special case of the general one in your question. – Julian Knight Jul 19 '12 at 21:11
  • @Synetech: That's the point of hardlinks -- both are the "actual file". There's no concept of an "original file" when it has multiple hardlinks. You simply put the file in both places. – afrazier Jul 20 '12 at 13:54
  • @afrazier, not really, when you download a file, you still store it in a directory somewhere before creating hard-links to it. Windows’ implementation is getting there, but it’s not quite like in Linux. Be glad that you don’t have OCD and “little, insignificant details” like this don’t bother you. – Synetech Jul 20 '12 at 13:59
  • @Synetech: After you've hardlinked it to its final destinations, delete the hardlink in your donwload folder. Or move the file from the download folder to one destination and hardlink to the others. Both actions leave you with the exact same result on disk: 1 file with more than 1 hardlink. – afrazier Jul 20 '12 at 15:09
  • > Or move the file from the download folder to one destination and hardlink to the others   Which one? OCD! As far as I’m concerned, hard-links are not a be-all—end-all solution to this; they’re just duct-tape. – Synetech Jul 20 '12 at 15:11
  • @Synetech: It doesn't matter. A file with multiple hardlinks is not a file with multiple pointers that say "the real file is in this folder", it's multiple directory entries to the same data extents on disk. Each link is indistinguishable from the others -- there's no "first" or "original". – afrazier Jul 20 '12 at 16:07
  • Like I said, hard-links only help with the situation; they are not the be-all–end-all solution. Having a file appear in both locations is not quite right. It doesn’t matter if there is only one copy on disk, it still exists twice instead of existing once and having multiple properties. The concept of a file-system needs to be completely overhauled to accommodate today’s needs. If hard-links are good enough for you, then by all means go ahead and use them; but they are not the final, best solution to tagging, sorting, searching, and organizing files. – Synetech Jul 20 '12 at 16:15
  • let us continue this discussion in chat – afrazier Jul 20 '12 at 17:37
  • The link for the duplicate doesn't exist anymore... – jiggunjer Sep 21 '16 at 10:48

1 Answers1

0

There are certainly tools like this.

The better document management systems allow for files to be uploaded to a folder structure and will sort out the meta-data required.

There are non-general tools as well. Check out Google's "Picasa" picture management tool which works as you've described. It monitors a set of folders for new images and automatically indexes and cross-references the meta-data. Any changes to meta-data are written back to the files in a standard way using EXIF and IPTC meta-data standards so that the files remain fully portable.

Your implication though is that you want to do this without the use of tools. This is not possible since non of the OS's are built this way natively. You would as a minimum need to use a replacement "explorer" interface that understood how the meta-data was being indexed and managed - hence the need for a document management system. The other issue is that decent DMS's don't come cheap. Of those that do have free versions, the majority will now be web browser based - such as Mayan and these are rarely as comfortable to use as a native interface.

Julian Knight
  • 14,501
  • 3
  • 29
  • 44