When I check a text field that depends on the value of another text field, instead of highlighting, as you mentioned, I intercept the change in value and accept or reject the change.
I found that if I try to change the text box B based on the change in the text box, then AI always runs into problems, but if I accept or reject the change in text box A based on the value in the text box B, it works fine.
Edit:
I assume that you are dealing with text fields for numbers only. If this is the case, I suggest you do what I did. First, for my validation rule, I created a polygon that represents numerical constraints, this allows you to have a linear relationship between the two values. Then, when a single value changes, I check if the current data point intersects with the check polygon. If the data point is outside the polygon, you do not allow this.
The problem you are facing is that if you are on the edge of the polygon, the value seems to be βstuckβ. To get around this, I created a graphic that shows the test polygon with the current point in it. Thus, if the user presses one value to the limit, they can understand why they can no longer change.
For good measure, I added sliders along each axis of the graph so that they can easily change the value, and I added drag and drop functionality to the point.
Thus, you have a stable check, and it is very clear to the user how to change the values ββand why they cannot go beyond certain restrictions.
source share