Talk:Versioning file system

no modern?

So why is there no modern versioning file system, or is there? It always seemed like a nice feature to me when using VMS back in the stone age. Spalding 12:41, 18 June 2006 (UTC)

Vista does support it, and Norton GoBack also as an addon to Window$. This is indeed a great feature and I don't know why isn't it more widespread  doc Aberdeen ❖ 01:17, 27 May 2007 (UTC)

Vista isn't mentioned in the article, however. Would someone mind adding it? – 80.212.184.71 (talk) 21:40, 14 December 2009 (UTC)

What about "mod_dav_svn"? It is sort of an adapter between the versioning system Subversion and WebDAV. WebDAV in turn is a network protocol for filesystem access, and I guess it can even be mounted as ordinary filesystem in some operating systems. Altogether, this solution might be concidered a modern versioning file system. Can someone verify this, and change the article accordingly? Michael R. 14:25, 17 October 2006 (UTC)

NTFS

What about NTFS's versioning through file's alternate datastreams? Shouldn't it be considered as a kind of versioning file system?

I mentioned NTFS, although it would be surprising if it was done via alternate streams. (Doc aberdeen 00:33, 27 May 2007 (UTC))

Surprisingly enough, it is. And it doesn't show up on the file size (which is just the size of the unnamed stream) or on the directory size.

Ummmm, no it isn't. NTFS does not keep versions of the files that it manages. The closest thing that some Windows boxes have for versioning is Shadow Copies, and that uses a separate cache file (or hardware), and not alternate datastreams. If you're talking about the "Version" tab of the File → Properties dialog in Explorer, please realize that that has nothing to do with Versioning File Systems.

TOPS20

I'm fairly certain that the file system used in ITS, which started development on the PDP-6, is far older than TOPS20. It's a bit of an oddity, though, since it don't have to be a versioning file system, but can be if you give your files a number (which will then be incremented automatically). Does that feature disqualify ITS? Otherwise I suggest it was the oldest file system with versioning, unless someone can tell us that WAITS or some other oddity was earlier. —Preceding unsigned comment added by 130.15.24.166 (talk) 22:52, 7 September 2007 (UTC)

Agreed, ITS counted as a versioning file system, and predated TENEX (which became TOPS-20). I am changing the article. ScottBurson (talk) 05:57, 25 January 2009 (UTC)

A slight clarification: ITS dates to 1967, and TENEX to 1969, so ITS is not "far older". Still, it's a bit older. On the other hand, I do not know for a fact that ITS had versioning from the beginning; it could have borrowed it from TENEX, AFAIK. I have edited the page to suggest that ITS was probably first, but not to claim this outright. ScottBurson (talk) 07:06, 25 January 2009 (UTC)

Removing information

I am removing information related to the following:

• Shadow Copy
• Time Machine
• TimeVault

It says in their descriptions that they are not versioning file systems, so they are not relevant to the article. Gh5046 (talk) 15:54, 23 January 2008 (UTC)

No, they are not versioning filesystems, although their GUI presentations may give this impression. In fact, they are all essentially front-ends for the rsync command:

rsync -aP --link-dest=PATHTO/$PREVIOUSBACKUP $SOURCE $CURRENTBACKUP

I'd recommend giving some indication of ho many of these filesystems are using straight copy-on-write, and mentioning that rsync derivative backup solutions use hard links for precisely this function. 78.148.242.178 (talk) 20:07, 8 March 2009 (UTC)

btrfs

Is btrfs really a versioning file system? It doesn't say so on its page. --91.67.244.235 (talk) 14:13, 13 April 2009 (UTC)

Agreed. I haven't seen anything like that. I marked it and will remove it soon if a reference is not added. (It *does* support snapshots, but that's different in both effect and classification on the page, by which it should be in "related software".) 128.105.181.52 (talk) 23:16, 3 February 2012 (UTC)

Btrfs supports copy-on-write cloning. From the btrfs page: "File cloning (copy-on-write on individual files, or byte ranges thereof)" Copy-on-write cloning is what versioning does, so wouldn't this count? Cbcoleman (talk) 22:20, 6 June 2012 (UTC)

copy-on-write is an optimization, though it could be used in a versioning filesystem. I btrfs isn't a versioning filesystem last I looked. When I looked last it was using COW for file copy optimization (as probably other actions). It saves space for copies that aren't changed, and for ones that are they have a lower io overhead than on normal filesystems (only content with the new changes are written, other systems write the whole file and then write the changes). — Preceding unsigned comment added by 64.73.243.146 (talk) 18:25, 3 October 2012 (UTC)

Btrfs is not is not a versioning filesystem. There is no mention of versioning in the btrfs main website; copy-on-write is not the same (as stated above, it is an optimization), and snapshots are also different because they are discrete actions (the administrator has to request a snapshot). Versioning requires that the action be automatically part of the file saving process. Since this has been disputed since 2009 with the most recent update in 2012, I am deleting it from the main page. — Preceding unsigned comment added by Wsloand (talk • contribs) 11:32, 24 March 2014 (UTC)

It appears that Btrfs supports "snapshots" which can be used for versioning: http://www.oracle.com/technetwork/articles/servers-storage-admin/advanced-btrfs-1734952.html "Snapshots, Clones, and Seed Devices" "snapshots" sub bullet. I would consider this versioning, and propose its readdition if no complaint here in a week or so (probably under related software since it seems to have the same bullet point worth as ZFS). Cheers! Rogerdpack (talk) 15:58, 25 August 2015 (UTC)

dropbox

I'm removing dropbox reference because it is just an application that keep file version on server (if you pay for it). It works like git, mercury, subversion, cvs etc and it is not a "similar functionality" — Preceding unsigned comment added by 189.27.8.52 (talk) 06:31, 5 October 2012 (UTC)