JOhnH
Touchdown! Greaser!
I use a commercial program for our veterinary hospital. Since we went to all Electronic Medical Records in 2007 we do not retain any paper. We scan everything in and attach it to a client/patient record.
I remember in my old tech support days (VAX/VMS) that we always had performance problems when we created a directory with too many files in it. So I arranged to store all of our attached files in a directory tree as such:
C:\top directory\scanned files\yyyy\mmm\long-descriptive-filename.
the long descriptive file name is formatted as:
\client# - patient name - descriptioin
(the patient name is the animal's name).
This way, the largest single directory I have has about 900 entries. It happens to be C:\top directory\scanned files\2013\may\
But my vendor insists that I convert to their scheme before I upgrade to the next version of their software. They want:
C:\application\client#\patient name\filename.
But this way, the subdirectory "client #" would have about 15,000 entries now and would continue to grow as I add clients.
They say my way is "too deep" (5 levels) vs (4 levels)
I say theirs is "too fat". 15,000 entries in a dir vs 1,000 entries.
The application runs on 32bit XP Pro workstations but the data is contained in "flat files" on a 64 bit server.
Does this make sense to anyone?
Would I gain or lose performance if I change?
They say I can choose either way but they don't support my way. And once you do anything that is "not supported", it aint fun to get past that with the script reading phone support to work on other problems.
I remember in my old tech support days (VAX/VMS) that we always had performance problems when we created a directory with too many files in it. So I arranged to store all of our attached files in a directory tree as such:
C:\top directory\scanned files\yyyy\mmm\long-descriptive-filename.
the long descriptive file name is formatted as:
\client# - patient name - descriptioin
(the patient name is the animal's name).
This way, the largest single directory I have has about 900 entries. It happens to be C:\top directory\scanned files\2013\may\
But my vendor insists that I convert to their scheme before I upgrade to the next version of their software. They want:
C:\application\client#\patient name\filename.
But this way, the subdirectory "client #" would have about 15,000 entries now and would continue to grow as I add clients.
They say my way is "too deep" (5 levels) vs (4 levels)
I say theirs is "too fat". 15,000 entries in a dir vs 1,000 entries.
The application runs on 32bit XP Pro workstations but the data is contained in "flat files" on a 64 bit server.
Does this make sense to anyone?
Would I gain or lose performance if I change?
They say I can choose either way but they don't support my way. And once you do anything that is "not supported", it aint fun to get past that with the script reading phone support to work on other problems.