I have spent several days now struggling with the coding of the new translator editor, and the result does not look good. My problem arises from the multitude of TERecord boxes littered over the screen. Imagine a set of boxes, some text variables and some TERecord boxes, arranged in sequence. The storybuilder decides to insert some text in a TERecord box. The TERecord box is only as large as the current text. Thus, when the storybuilder types a character, that TERecord box must be increased in size and redrawn. But this implies that all the downstream graphical structures must be repositioned and the screen redrawn &emdash; all for a single keystroke. This will not work. It will be too slow to type. It's time to abandon this snazzy idea and revert to the older system with a regular text editor and control codes. Dirty ricklefricks.
So we're going to have one TERecord for each POV. We'll use the same old conventions for apposition, text variables, etc. I won't be able to highlight text variables or appositives, because the TERecord must be internally homogeneous. I can add a spellchecker that corrects misspelled text variables, and a grammar checker to insure that all brackets are balanced, etc. Small consolation.
OK, back to work, Mr. Genius.