Subject: | |
From: | |
Reply To: | |
Date: | Mon, 19 Feb 2007 16:03:02 -0500 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
John,
I think you are confusing application level meta data with operating system
level statistical data.
When a file is created, updated and in some cases read the operating system
updates the file statistics. In this case the operating system does not
know or care what application the file belongs to. In fact if it did this
would probably be controversial in light of anti-trust law.
In some operating systems such as MS Windows you can change the behavior of
the OS using registry "tweaks" so that you can turn off last read
statistics to enhance efficiency. It would also not be hard to write a
program that changed any of the statistics on mass (if you wanted).
On the other hand you have Application Level meta data, this is created and
maintained by the application (Office, Acrobat et al). Copying or moving a
file does not invoke these applications so this meta data will not be
changed.
It is common for applications to hook into the Windows system to add
functionality. So in the case of Adobe Acrobat, it installs a hook which
adds a PDF tab when you right hand click on a file. The code in this tab
would open the file and extract data for use in the PDF tab. The process of
reading this file will update the last accessed statistic at the operating
system level (if this has not been disabled).
The functionality you seek has been provided on Mainframes via SMF for
decades. This functionality has slowly been creeping into Windows and
should be more advanced in Longhorn when it is released mid-year.
As for UNIX/Linux audit capabilities are sadly not at the forefront of
development activities.
On the issue of courts understanding the ins and outs of technology, this
is all relatively new to them. In most cases arguments related to matters
such as these are simply an adjunct to the case. Any finding should
therefore be seen with a view that they will evolve as precedents are
challenged and new arguments emerge.
Gerard
List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance
|
|
|