in reply to Re^3: time difference
in thread time difference

The Unix founders argue (and I agree) that "creation time" is at best an odd concept.

Care to explain that? 

dir thisfile
File not found.
echo . > thisfile
dir thisfile
01/08/2005  06:38                 4 thisfile

[download]

The file was created. The time tells you when it was created. How is that false or ambiguous?

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
Lingua non convalesco, consenesco et abolesco. -- Rule 1 has a caveat! -- Who broke the cabal?
"Science is about questioning the status quo. Questioning authority".
The "good enough" maybe good enough for the now, and perfection maybe unobtainable, but that should not preclude us from striving for perfection, when time, circumstance or desire allow.

Replies are listed 'Best First'.
Re^5: time difference
by merlyn (Sage) on Aug 01, 2005 at 06:00 UTC
    The file was created. The time tells you when it was created. How is that false or ambiguous?
    Now, let's say I update the information.

    I could do it one of two ways. I can copy all the data to a new file, then rename the file back over to the existing name. Or I could rewrite the file in place with the new information: more dangerous, but just as valid.

    Should these both have the same original creation time, the creation time of when I started to do the update, or the creation time of when I finished closing the file?

    Should both methods result in the same creation time?

    See, the problem is, what does "creation" mean once you can change things? The "creation" of the string "hello world" where "HELLO world" was formerly is "right now", is it not? So why should the creation time reflect an older timestamp than just now?

    From when do you measure the beginning of the existance of "this file", when the contents can be changed?

    The Unix founders therefore argued that there is no sensible consistent definition for "creation time", and got rid of it entirely. Yeay.

    Instead, we get three very useful, and clearly defined, timestamps. Time of last access (useful to find stale or unused files), time of last contents-modified (used by make), and time of last contents-or-meta-info-changed (used by backups).

    -- Randal L. Schwartz, Perl hacker
    Be sure to read my standard disclaimer if this is a reply.

[reply]

      Okay. Thanks for the history lesson. :)

      I guess you could argue that there is a difference between a file and it's contents, and that a file moved or updated in place is a modification of the original file, whereas a file copied or overwritten is a different file, but I don't expect that would have much of an effect.

      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      Lingua non convalesco, consenesco et abolesco. -- Rule 1 has a caveat! -- Who broke the cabal?
      "Science is about questioning the status quo. Questioning authority".
      The "good enough" maybe good enough for the now, and perfection maybe unobtainable, but that should not preclude us from striving for perfection, when time, circumstance or desire allow.
[reply]

In Section Seekers of Perl Wisdom