Originally posted by br.davidson Could you explain your DAM strategy and why it's more efficient for the originals and derivatives to be stored together? I keep mine together right now too, but I'm seeing limitations to that strategy.
It's a matter of software, I think. The program I have has a versioning mechanism built into it as does Lightroom. The software is flexible enough that derivatives can be stored anywhere if you tell it that. However, if you have the derivatives in the same folder, you can tell it that is the setup and it will save itself from looking in all possible folders. The result is that when it looks for derivatives, it only has to look through a limited group of photos instead of all photos. In most operations having derivatives elsewhere wouldn't matter too much, but everytime files are added or updated, the software checks to see whether the new files are versions or originals to other files. It's a bit faster if the derivatives are kept in the same folder... Actually, the software is still fast if you keep derivatives in subfolders, too.
I believe LR works similarly, although I don't use it for DAM, so I'm not sure exactly.
More technically, my original images all have a similar file name scheme regardless of the camera and extension. They tend to have something like YYYYMMDD_CAMID_####.EXT The #### is the file number from the original camera file name and the CAMID is a fixed 5 character section to identify the camera it came from and the date comes from the EXIF date time original field. The derivative files are similar but the end of the file name has some ####-V##.EXT type field where V is actually a variety of characters to help identify the version (A = typical shot, B = black and white, C = heavily cropped, H = Hdr, P = Panorama, etc). As a result, I can instantly find all photos that are versions by the "-V01" file name and originals don't have that. My DAM software could also go by date and time, but I find that to be a bit slow and problem ridden as some software can change EXIF date and times, and reading EXIF for each and every file is a little more computational power than just reading the file names.
I do see why separating the files could be important to your setup. I don't use a cloud service, so that may make me rethink my system. That was why I made the earlier reply about being ABLE to separate originals from derivatives. In my case, I could use a file renaming utility that can handle regular expressions to select all files with a "-V00" type pattern and move them into a new folder scheme. You are essentially looking to use the same folders as before but rather add one level at the top. A good advanced file renamer should be able to grab all of your jpeg files and "rename them" such as
JPG\Orig_Folder_Tree\File_Name.jpg
You just need to find one that works and put a little bit of time into making it work for you. It seems the more features in such a utility, the more difficult they are to use.