On 2014-10-04 10:05, John Andersen wrote:
On 10/3/2014 10:36 PM, Felix Miata wrote:
I do have a problem with the notion that both creation and last modified both exist as distinct "regular" file attributes. Where at filesystem level can relevance lie for preserving creation when last modified is newer?
Wait, What? This whole thread has been about wanting the preserve the date that photos were taken via the file-system date. Now you question the usefulness of that?
The fact that 'nix systems have always preserved both dates and Microsoft felt compelled to add it in NTFS as well as exFAT should suggest this isn't the only reason for doing so.
In fact, Linux filesystems store *three* different timestamps. Vfat keeps one. Other, more recent, Windows filesystems, keep two, creation and last modification, because it simply is useful. cer@Telcontar:~> stat p File: ‘p’ Size: 109 Blocks: 8 IO Block: 4096 regular file Device: 10301h/66305d Inode: 134977 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ cer) Gid: ( 100/ users) Access: 2014-10-02 23:19:09.904084937 +0200 Modify: 2014-07-14 05:10:37.000000000 +0200 Change: 2014-10-01 17:28:38.886944560 +0200 Birth: - cer@Telcontar:~> Notice that it is not creation date, but modification date. -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" at Telcontar)