Easy to understand and easy to parse and implement is what I consider important.
Well, parsing the X/Y in the Interactive Diagram was as trivial as it can get. After reading an atom and looking up the (x,y) step and default range, you have to examine the following characters anyway for being a digit that would overrule the range, or duplication to toggle the range 1<->inf. When you encounter an X in that process you just add (3,0) to (x,y) (and (2,2) for an Y), and keep looping until you encounter the digit or something else. Having to test for en entirely new (x,y) syntax would have been far more complex.
Betza notation is intrinsically more difficult to understand than (x,y) notation; you have to remember what all letters mean. If that was an argument you should not use Betza notation at all. And only having to remember that X boosts the leap by 3 steps gives you a lot of extra leaps for just remembering one more letter.
Well, parsing the X/Y in the Interactive Diagram was as trivial as it can get. After reading an atom and looking up the (x,y) step and default range, you have to examine the following characters anyway for being a digit that would overrule the range, or duplication to toggle the range 1<->inf. When you encounter an X in that process you just add (3,0) to (x,y) (and (2,2) for an Y), and keep looping until you encounter the digit or something else. Having to test for en entirely new (x,y) syntax would have been far more complex.
Betza notation is intrinsically more difficult to understand than (x,y) notation; you have to remember what all letters mean. If that was an argument you should not use Betza notation at all. And only having to remember that X boosts the leap by 3 steps gives you a lot of extra leaps for just remembering one more letter.