I was wondering if we could have the following feature: embed the code (the bracketed string a user types) into the image generated (for some image formats that supports custom data embedding).
An example of this feature can be found in LaTeXiT. The LaTeX code is embedded into the image if the output format is chosen as e.g., PDF.
This frees the users from having to save the original code of the tree. If a user wants to edit the tree, they can simply recover the code by reading the data from the image (This will need to be accompanied by another new feature that decodes the image, of course).