Retighten specification of NXdata's axes attribute for #1381 #1392
Conversation
Tweak examples and clarify that alternative choices for axis values can be indicated by the presence of its corresponding AXISNAME_indices attribute
|
@jacobfilik, please review too. |
There was a problem hiding this comment.
As I mentioned in #1381 I do agree we approved #1213 without voting while it does change the standard. So we should put it back.
But I don't think this tweak is enough, that's how we ended up with an inconsistent and impossible to read NXdata doc in the first place.
I'll try to make an alternative PR that explicitly states that the length of @axes must be equal to rank of signal. And then discuss how you need to find all axes of NXdata and change the examples to show what to do with the inconsistent definition as a result of the constraint.
|
|
Abandoned in favour of #1396 |
Tweak examples and clarify that alternative choices for axis values can be indicated by the presence of its corresponding AXISNAME_indices attribute