SIFT - desired enhancement

Option to exclude DESCRIPT.ION files (actually, %_dname). Those of us who use TCC to back up to non-NTFS devices must use the file, not the stream form of descriptions. The timestamp of these files is that of the last time a file with description was backed up, and it can dominate a recursive SIFT. A more extensive enhancement would be the full implementation of the file exclusion range. Another alternate to achieve the same basic goal (but which may falsify the results too much) would be an option to ignore hidden files.

Charles Dye

Super Moderator
Staff member
May 20, 2008
Albuquerque, NM
On second thought, I've already implemented ranges once, so the job is largely cut-and-paste.

I've uploaded another build which theoretically supports ranges. Steve, I think you just volunteered to beta-test it :)
Sorry for such a looooooooong delayed report - I tested /nj, date, size, attribute, and exclusion range; everything works as on internal commands. I had come back to the thread to ask for the /nj feature, but it is already working. A few more possible enhancements:
1/ option to use an invisible array (one not reported by SETARRAY) simplifying the cimmand line (array need not survive command)
2/ a variable for default options similar to VARIABLEEXCLUDE (aliases are a viable option negating this)
3/ quiet mode - nothing to screen (if conflicting /v is also specified, last one wins)
4/ for /V mode choice of what is displayed
- header
- footer
- which column(s)
5/ summary variables in the style %_sift_XXX:
number of files reported (relevant if less than requested)
number of files and directories scanned
limit parameter value, i.e., the size of the largest or smallest file when /L or /S are used, etc.
(alternately all 4 values: smallest size, largest size, oldest timestamp, newest timestamp)
The limit is very useful if one cares only for one or more of these. I use it often in FTP transfers to avoid repeated downloading.

Thanks for the very useful work!