Like Peter Clifton, I worried that examining the contents of all
candidate files would be slow. So I did just implement a quick
test of that. It opened every file, and searched for an Element
keyword.
On my machine, a library of 11000 files took 114 seconds to scan
when run immediately after a reboot. The second run took under
one second, because of OS caching no doubt.
Maybe that is an atypically large library, nevertheless the direct
approach appears to not be feasible.
Like Peter Clifton, I worried that examining the contents of all
candidate files would be slow. So I did just implement a quick
test of that. It opened every file, and searched for an Element
keyword.
On my machine, a library of 11000 files took 114 seconds to scan
when run immediately after a reboot. The second run took under
one second, because of OS caching no doubt.
Maybe that is an atypically large library, nevertheless the direct
approach appears to not be feasible.