http://brainstorm.ubuntu.com/item/9560/
If you use a hierarchical filesystem, you have to duplicate your file when you want to put them in the folders or making symbolic links which make it difficult if you want to delete it.

A tag filesystem on the other hand solve completely the problem. And because it can be implemented as kernel filesystem, every application can then use it instead of being an extra application.

Because of its interest in everyday use of the computer I put this project in the category office

Please check here:
http://www.tagsistant.net/

some articles:
http://nascent.freeshell.org/programming/TagFS/
http://citeseer.ist.psu.edu/741340.html

also some interesting work:
http://code.google.com/p/tagfs/


similar ideas have already been posted:
http://brainstorm.ubuntu.com/idea/1981/
http://brainstorm.ubuntu.com/idea/1982/
http://brainstorm.ubuntu.com/idea/3798/


[67 votes] Solution #1: Auto-generated solution of idea #9560


]]> en-us Fri, 06 Jun 2008 07:40:32 +0000 Sun, 29 Jan 2012 14:33:58 +0000 QAPoll module http://brainstorm.ubuntu.com/idea/9560/ Fri, 06 Jun 2008 09:15:04 +0000 Fri, 06 Jun 2008 18:01:06 +0000 Nepomuk seems to be a good project
but what about all other applications including the so terminal that do not include nepomuk

Also it probably makes things slow.

nevermind85
hard links have do be manually set (for each tag/folder you want to put it) and if you want to remove a file for all tags/folder, you will have to go to all tag/folder to delete it.

]]> Fri, 06 Jun 2008 22:05:19 +0000
One good tag to place on files is the current task. But this would require knowing what project the user is working on. With a task-oriented UI, you would have that information.]]> Sat, 07 Jun 2008 05:15:39 +0000
Stratus - [url=http://wealthandpower.org/stratus/]http://wealthandpower.org/stratus/[/url]

FITaFS - [url=http://fi-mihej-en.blogspot.com/]http://fi-mihej-en.blogspot.com/[/url] (now). My development. There is while little information because I will lay out complete description hardly later - after a publication in a paper magazine (in a flow month-two). Approximately in also plan to lay out time the first working version of the module for FUSE. Why did I decide to create another tag-based filesystem? One of reasons are serious problems, showing up at other tag-based filesystems, at presence of plenty of tags: 1) to the man becomes uncomfortably in them to be oriented; 2) at the programs, trying to go round all tree folders failures appear.

P. S.: I am sorry for my English.]]> Tue, 10 Jun 2008 15:11:44 +0000
FITaFS - http://fi-mihej-en.blogspot.com/

sorry ]]> Tue, 10 Jun 2008 15:15:58 +0000 Mon, 25 Aug 2008 16:52:07 +0000
Tagsistant is developed with ease of use for the average user as one of its primary goals. Some design choices that may sound unnatural at first glance are a direct consequence of the research of a intuitive interface both at shell prompt and inside a file manager.

I hope Ubuntu can get interested in it and someday will plan the integration.

Best wishes.]]> Tue, 09 Sep 2008 08:41:37 +0000
I've never liked the idea of categorizing files based on their file type, like with Microsoft's "My Pictures" or "My Music" folders. They should be categorized based on their content.

If you go on a fishing trip with your family, do you really want the photos to be in a Photos folder, the videos to be in a Videos folder, and maps and emails in completely different folders? Of course not. You want everything related in the "2008 Family Fishing Trip" folder.

With a tag-based system, they would automatically be tagged with metadata like the date they were taken: "date:2008-10-12", and you could also add descriptive tags like "family", "fishing trip", and "vacation". The images would automatically be tagged with something like "mime:image/jpeg", and the act of downloading them from your camera would automatically tag them with "Photos", to differentiate from other types of graphics that can be stored in jpegs. A map of the directions to the fishing hole, for instance, would have a "Maps" tag instead of "Photos", but would otherwise have all the other tags, so it would show alongside the photos if you searched for "2008 family fishing trip". The photos would be tagged with the name/ID of the person who took them, and then you could select a bunch of images and tag them with the names of the people that are in them.

Then you could navigate and find things by topic, by date, by file type, or any combination! You could type a friend's name and the word "photo" and instantly see every picture you own of them at once, or you could type the name of an event and view all the photos and other related files for a single event at once. Or you could type a name and an event and see all the pictures of them at that particular event, as well as emails to them about the event, etc.]]> Wed, 15 Oct 2008 04:53:36 +0000
]]> Wed, 15 Oct 2008 05:18:47 +0000 "what about hardlinks?"

Hmmm... Are links supported by all filesystems? Maybe tags could be implemented as hard or soft links. The tags exist as folders (maybe hidden folders?) in the Home directory, and files would be linked into each folder if they were tagged. Then file browsers would interpret this with a more advanced interface.

The best solution would be independent of desktop and independent of file system.]]> Mon, 20 Oct 2008 14:34:48 +0000
Surely this has been proposed before, though. What are the proposed implementations?

I imagine a system in which different "kinds" of properties of the file are encoded in the tag, and the human-readable GUI implementation of the tag might show different classes of tags ("names" given by the user vs file types generated by the system, for instance) in different colors, maybe. So if you searched for "jpeg" and "Family", the jpeg would be recognized as a file type and shown with some kind of filetype prefix and color, and the "Family" would be recognized as a user-generated label and given a different color?

This looks relevant:
http://www.foo.be/cgi-bin/wiki.pl/MachineTag
http://www.flickr.com/groups/api/discuss/72157594497877875/


]]> Mon, 20 Oct 2008 15:29:59 +0000 Wed, 22 Oct 2008 21:16:54 +0000 Wed, 29 Oct 2008 20:19:21 +0000
For example:

cd /usr/share/hello-world

would work... but so would:

cd /hello-world/share/usr

or simply

cd /hello-world

(assuming there is only one hello-world in the filesystem)]]> Tue, 04 Aug 2009 20:18:38 +0000 Also I think that Metadata should be reformed. All metadata is currently done through different ways, usually information is just stored about the name of the file, the time the file was created, and the permissions.

MP3s use ID3 tags to store track information. This leads to problems with formating. But what if ID3 tags were done through the filesystem level? What if there was a tag for each author just like there would be for the time, or the name, this would end a lot of compatibility issues.]]> Fri, 16 Oct 2009 18:23:02 +0000
http://code.google.com/p/tagfilesystem/
]]> Fri, 08 Jan 2010 05:54:32 +0000
http://pages.stern.nyu.edu/~marriaga/software/oyepa

None of the FUSE complication and stuff. my only regret is that there's no ubuntu packages to install from.]]> Fri, 19 Feb 2010 02:20:23 +0000
I think specially on my music library. A mp3 file has logical properties including Filename, Mimetype (audio/mpeg), Title, Artist, Album, Genre, Year, Bitrate, and Duration.


== Virtual Folders

There have to be virtual folders to aggregate specific mime-types. There may be virtual folders like 'music' (mp3, ogg, wav, etc.), 'videos' (.mpg, .m4v, mkv, etc.), 'documents' (.txt, .pdf, .doc, etc.), and images (.jpeg, .png, etc.).


== Sorage on the native File System

The location where the files are stored is probably a flat directory, while the meta-information will be stored in a (mysql) database. The filename is probably a hashcode of the file to prevent name conflicts, so the filename has to be stored in the database.

Export has to be possible to. The metainformation has to be written back into the file and the file has to be renamed on export.


== Example Queries

Query of the database might look like:

> ~$ mp3player -add "/fs/mime:audio*/artist:Lady Sovereign/Album:Public Warning/"

or simple:

> ~$ mp3player -add "/fs/music/Lady Souvereign/Public Warning/"

The will add the same files as the command

> ~$ mp3player -add "/fs/Public Warning/Lady Souvereign/music/"


== Comparism to Tag based File System

This looks like a tag based system to the user, but in the background works a triplestore (or quadruplestore).

The meaning can be reasoned by the application. 'music' is an alias for 'mime:audio*' and 'Lady Souvereign' might return anything that has to do with Lady Souvereign, but will restricted to music because of the first expression.

The expression '/fs/author:Lady Sovereign/' will only return books written by Lady Sovereign, while '/fs/artist:Lady Sovereign' will only result songs. The tag-like expression '/fs/Lady Sovereign/' will return everything that has to do with her: Books from her, Books about her, Songs from her, Remixes not from her, Videos showing her, and the like.

== Advanced Queries

There might be the possibility for an advanced query syntax. Examples might be AND, OR, and NOT. They might be seen as alias for set-operators: "set:union", "set:intersect" and "set:complement".

Example:

> ~$ mp3player -add "/fs/[artist:LadySovereign/album:Public Warning]/set:union/[artist:Michael Jackson/album:Thriller]/"

or short:

> ~$ mp3player -add "/fs/[LadySovereign/Public Warning]/AND/[Michael Jackson/Thriller]/"

This command adds songs from Lady Souverein as well as Michael Jackson to the playlist. Note that i use Square Brackets to clarify the logical order of processing.


Special expressions might be possible for comparism operators like more-then, less-then, less-then-or-equal, etc.:

> ~$ mp3player -add "/fs/music/[duration:ge:120]"]]> Tue, 21 Dec 2010 01:11:36 +0000
It can be used standalone to tag files already on a filesystem, but also as a Fuse filesystem, using the traditional hierarchical commands (mkdir, mv, ln, ls, ...) or even in graphical file managers as nautilus

If you want to try labelfs:
http://code.google.com/p/labelfs/]]> Sun, 10 Apr 2011 18:53:32 +0000 Sun, 29 May 2011 17:36:58 +0000 I'm an ubuntu fan but not that much of an expert with programming, as my job is to teach italian language and culture.
The reason why I'm writing is related to my profession:
I've been thinking for a while now that it'd be really cool if I could scan my language textbooks and my grammars in order to have them all digitalized and always available without having to carry around tons of books.
The next step in my dream would be, after scanning, be able to add tags to the documents. My understanding, so far, though, is that I can only add tags to the whole file. So let's say I have a pdf with an italian grammar in it, I could only tag the grammar itself, and not part of it.
What I wanted to do is, instead, tagging "areas" of the pages, adding tags such as "past principle", or "articles in italian", and so on. By doing so in many different textbooks, whenever I look up "past participle" I would get as a result links to the exact page (or list of pages) where that topic is dealt with in all the textbooks of my library. Did I kind of explain myself? Is there a way for me to do that?
Pdfs are good because they are light, but so far I couldn't find a solution. A different idea would be put single scanned pages, scanned as images, inside a directory so to have a single file for every page. Kind of messy, though.
Any ideas? Problem is my very (very! almost nothing...) limited understanding of programming and system architecture also limit my chances to imagine a possible solution for the problem...hopefully you'll be able to help me with that!
Thank you for now!]]> Fri, 17 Jun 2011 10:44:38 +0000 Fri, 19 Aug 2011 11:31:34 +0000 juegos de autos gratis]]> Mon, 02 Jan 2012 19:54:40 +0000 juegos de auto gratis juego de auto gratis]]> Thu, 05 Jan 2012 23:02:16 +0000 funbrain]]> Sun, 08 Jan 2012 17:56:15 +0000
click here]]> Tue, 10 Jan 2012 06:50:18 +0000 dubai property for sale | dubai marina property | Arabian Ranches Dubai


]]> Sun, 29 Jan 2012 14:33:58 +0000