from email from @abrasax
The simplest fix I can think of for the metadata issue would be to distinguish the temporary input box of the (potentially) multivalued authors or contributors from all the other boxes, but have the final collected values in a box that looks the same as the others.
eg- make the input boxes of those two characteristics dotted and the box in which the successive values accumulate appear identical to those of all the other values. If they want to edit one value they select it as they do now and it goes back into the dotted box for editing or deleting as it does now. I think that much of the counterintuitive nature of the current method comes from accumulation of the values between the lines- outside of the boxes that indicate choices made. I think the only change necessary would be to have two boxes- readily distinguishable from one another, with only the final one identical to those with single values.
from email from @abrasax
The simplest fix I can think of for the metadata issue would be to distinguish the temporary input box of the (potentially) multivalued authors or contributors from all the other boxes, but have the final collected values in a box that looks the same as the others.
eg- make the input boxes of those two characteristics dotted and the box in which the successive values accumulate appear identical to those of all the other values. If they want to edit one value they select it as they do now and it goes back into the dotted box for editing or deleting as it does now. I think that much of the counterintuitive nature of the current method comes from accumulation of the values between the lines- outside of the boxes that indicate choices made. I think the only change necessary would be to have two boxes- readily distinguishable from one another, with only the final one identical to those with single values.