Re: Re: File modification date

From: Francesco Pizzolante <fpz-djc/iPCCuDYQheJpep6IedvLeJWuRmrY@public.gmane.org>
To: "Sébastien Vauban"
	<wxhgmqzgwmuf-geNee64TY+gS+FvcfC7Uqw@public.gmane.org>
Cc: emacs-orgmode-mXXj517/zsQ@public.gmane.org
Subject: Re: Re: File modification date
Date: Mon, 07 Dec 2009 17:13:21 +0100	[thread overview]
Message-ID: <87bpiavk2m.fsf@missioncriticalit.com> (raw)
In-Reply-To: <877hsyzy0t.fsf-pwAqS3aGAJQybS5Ee8rs3A@public.gmane.org> ("Sébastien Vauban"'s message of "Mon, 07 Dec 2009 14:58:26 +0100")

Hi,

>> DATE was implemented exactly to specify a fixed date. You can get the
>> modification date with
>>
>>     {{{modification-time(%Y-%m-%d)}}}
>>
>> and the current date with
>>
>>     {{{date(%Y-%m-%d)}}}
>>
>> These will be expanded upon export.  So you can, in fact, do
>>
>>     #+DATE: {{{modification-time(%Y-%m-%d)}}}
>
> The problem is that the author does not see anymore the "last modification
> date", then, except in the PDF output.
>
> And, in fact, if that Org file is put under VC, you can have two people
> printing (exporting) the exact same document and it would still have a
> different date. That's because, by default, the original modification date is
> not kept on the file system when checking out a file from Subversion (my
> current VC system).
>
> To do so, one need to set to his =.subversion/config=:
>
> #+BEGIN_SRC sh
> # keep the modification time from the repository, for Makefiles, etc.
> use-commit-times = yes
> #+END_SRC
>
> So, I find it would be good for the author to see the "real" last modification
> date (which, moreover, can be different from the "copy" date of the file in
> your file system).

In the same line of thinking, it would be good to keep trace of the "real"
creation date too and, as you suggest it, by completely separating these dates
from the file system dates.

What matters here are the dates corresponding to the content of the file, not
the file itself. And those dates could be exported to the PDF as well (or
other types of exportations).

In order to clarify my point of view, I see 3 important dates to keep:

1) The creation date:

   It could be defined with

   #+CREATED: 2009-12-01 09:36  or  #+CREATION_DATE: 2009-12-01 09:36

   and, when exporting to LaTeX, it would be exported to

   \pdfinfo{/CreationDate (D:200912010936)}.

   It will then appear in the PDF properties of the exported document as

   Created: 2009/12/01 09:36:00.

   This date is defined once and never updated. It can even be added
   automatically by the C-c C-e t command when creating a new buffer.

2) The modification date:

   It could be defined with

   #+MODIFIED: 2009-12-07 16:45  or  #+MODIFICATION_DATE: 2009-12-07 16:45

   and, when exporting to LaTeX, it could be exported to

   \pdfinfo{/ModDate (D:200912071645)}.

   It will then appear in the PDF properties of the exported document as

   Modified: 2009/12/07 16:45:31.

   This is the date that you will see updated each time that you change and
   save your org buffer. It could be done automatically with org-mode.

3) The last date is the date you want to display in your document (in effect,
   the one that is exported as \date(XXX) in LaTeX).

   This date could be the last modification date (in this case, you would not
   need to specify it) or give it a fixed value (in this case, you specify it
   as it works now: #+DATE: 2009-12-09).

With these dates, a document would contain all the information regarding it's
life cycle independently of the file system. And, in addition, the information
would be propagated in the exported documents.

What do you all think?

Regards
Francesco

_______________________________________________
Emacs-orgmode mailing list
Please use `Reply All' to send replies to the list.
Emacs-orgmode-mXXj517/zsQ@public.gmane.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode