http://ubuntuforums.org/showthread.php?t=169551
Other filesystems like ZFS and btrfs utilize copy-on-write to virtually eliminate file fragmentation -- files are copied whenever edits are made to the file.
This makes btrfs suitable for virtual machine images.
And if you're on an ssd its likely doing wear leveling in a similar fashion.
Really the better way to do this than overwriting files in place is to have encrypted files and keys and just reencrypt with a random key you don't retain, or some other method I suppose. But the ability to rewrite to the same spot in the filesystem, for a copy on write filesystem, is rather ludicrous. Hesitate all you want, I prefer cow fs's for most tasks.
You could solve this by automatically shredding deleted stuff in the background, but it would probably kill SSDs and other storage devices which you don't want to write to too often.
You really consider being able to shred individual files a critical filesystem feature? Over, say, not having to configure static partition sizes, having snapshots, or being completely immune by design to large classes of data corruption (including all those that require an fsck-like tool to repair)?