<tied>: 'placement' vs. 'orientation'

[lemzwerg]

I wonder why there are both the 'placement' and the 'orientation' attributes for <tied>. What example did trigger this implementation? Does something like

<tied placement="above" orientation="under" ...>

exist in the real world? Even in complex chords it is strictly forbidden to ever have such a situation...

BTW, in the 'musicxml-samples' archive, all examples use 'orientation' for <tied>, while almost all examples for <slur> use 'placement', with a few exceptions – and I couldn't see a difference. I would be very grateful for more insights.

[lemzwerg]

Thanks to Myke's site I can quote Michael's reply from a question asked on the now defunct MusicXML forum – which is available via the WayBack machine! He wrote:

Sometimes it is easier for a program to export orientation, sometimes placement. Whether the slur or tie itself goes over or under is a matter of placement. Whether the tips of the slurs or ties go up or down is a matter of orientation.

Usually the two are correlated. So yes, if one attribute is present but not the other, it should usually work to infer placement from orientation and vice-versa.

And in a follow-up reply:

I don't have a repertoire example of a difference between placement and orientation offhand. They indeed may not occur, especially for ties. DTDs make it difficult to express that two different attributes should not be used together. We should probably add something to the MusicXML DTD and XSD to explain that if you use orientation, you need not use placement, and vice versa.