Option to exclude XMP:Orientation when propagating or copying XMP data

[joel23]

Regarding to this thread which - by the options which were used - IMHO is not a bug but an expected behavior, it would be nice when we are able to exclude XMP:Orientation when XMP data is copied or propagated.

[Mario]

How would that work if the EXIF orientation is not excluded?
Or, more precise, the EXIF orientation variable is automatically mapped to the XMP orientation variable when the metadata is read after the propagation process so the XMP orientation is always replaced with the contents of the EXIF orientation (MWG compliance must be on).

[joel23]

How would that work if the EXIF orientation is not excluded?

gar nicht. Not at all. Of course some still need to use the "EXIF Data without Orientation" option as well.

Or, more precise, the EXIF orientation variable is automatically mapped to the XMP orientation variable when the metadata is read after the propagation process so the XMP orientation is always replaced with the contents of the EXIF orientation (MWG compliance must be on).

I know about this Mario, because this is what happened to Ferdinand. But I can imagine another option analogue to "EXIF Data without orientation".

Or even better throw that and create a new one or rename it to  "Keep Orientation" with something like "--exif:orientation --xmp:orientation" behind it. Regardless if some propagates/copies EXIF or XMP this would make sure that EXIF <-> XMP is not resynced by the other.

In the context of the mentioned thread Ferdinands propagation rule would look like:
"EXIF Data *"
"XMP Data without LR/CRS data "
"Keep Orientation"

Same for the copying dialog.

[Ferdinand]

My view is that they need to be linked somehow.  That is, not propagating exif-orientation should also imply not propagating xmp-orientation, and visa-versa.  I'm not sure of the best way to do this.

One issue I have with the propagation options is that they are independent options, i.e. you select them independently.  Yet there are some interactions between them, like this.  I can't see an easy way to deal with it. 

You said in the past that you didn't want to make this list of propagation options too long.  But I wonder if one way to address the issue would be to have sets of things to propagate.

As previously stated, I can't see a good reason for ever propagating orientation and I would exclude it from all propagation options.  Ditto for embedded colour profile.

[Carlo Didier]

I fully agree with Ferdinand on this point. It's something that bugged me from the start.

[Mario]

Done. Added an additional "Don't copy XMP orientation" rule set. This can be combined with all other XMP rule sets to explicitly exclude the XMP orientation.