+ - What ever happened to versioning file systems? 3
Submitted
by
shovas
shovas writes "As a developer and sysadmin, the benefits of revision control systems are clear. It only seems natural that a simpler, transparent approach to versioning files on a regular file system would be a net win. There's ext3cow and Wayback FS, and possibly some fuse-based projects, but each is either dead, immature or just not applicable. So what happened to the promise of versioning file systems? Hasn't everyone lost a file to a bad rm command? Hasn't everyone wished they could see a revision of a file in the past? What's the hold up?"
Competition from "snapshotting" filesystems (Score:1)
Re: (Score:2)
Must be hard to write an efficient diffing algorithm that works for most file types
What has a diff algorithm got to do with a versioning file system? Every time a file is saved, it should be written as a new version. Diff would be used to compare two versions of a file, not to decide whether to make a new version.
An ancient approach was used in DEC's FILES-11 system back in the 80's, where each file name had a version number attached (such as FILE.EXT;158). Files could be specified explicitly including the version number, but if a file was specified without a version number, then the h
Re: (Score:2)
http://en.wikipedia.org/wiki/OpenVMS [wikipedia.org]