@Joanna I am sorry if you find my tests and explanations too long and tortuous to follow I must work on that. The truth is that I find the same with some of yours and others where the writer knows and understands the problem but the reader can’t necessarily get up to speed quickly enough! But the cryptic posts don’t work either so …
But your solution is, I believe,
-
Largely unnecessary
-
Way too destructive, particularly to achieve harmony between a field laptop and a desktop system.
-
My alternative procedure that I espoused some time ago should work, although the deletion bug I have just found might just “interfere” with that.
-
The tests above worked perfectly and so I moved into testing with metadata in tow (in the containers that should be used) and attempted the tests after an afternoon chopping down vegetation and clearing up the debris with only a head torch for light! I made too many mistakes setting up the tests because I was tired and found some old bugs I have encountered before and the new “slight of hand” deletion bug.
-
None of this is really hard, the hardest part is actually understanding what is possible when we are working in a vacuum of non-communication from DxO.
I believed the fallacy about how hard hierarchical keys are and how they should be avoided at all costs! Then I realised it was simply a formatting issue and every package has adopted its own variation in rules! If you implemented my Keyword Formatting Template design your program could be everything to everyone on the MAC, except a purist like you!?
But that is actually just another entry in the table so your package can interwork between all the other packages!?
The problem with your package and the Python KFT script I am (slowly) developing when not gardening, DIYing, testing and writing huge posts, is that they are yet another component when the technology could so easily be built into DxPL @Musashi!?
The hardest part about the tests that I am currently doing is the time it takes to set them up and capture results when the tests are over in a couple of minutes!
So cryptic version of the long post
- According to my tests it is possible to synchronise two copies of the database, and images and DOPs between two system running identical versions of DxPL when the metadata is passed via the DOP. So well that DxPL can be active on both systems at the same time and the edit data will happily flip flop between the systems automatically!
What I was starting to test when I made errors and then found errors was to repeat that success with “conventionally held” metadata in tow!