It’s the OCR part that’s difficult. Most drawing packages support a tracing paper approach in which an underlying image is used as a guide to manually do what’s needed. Works well for simple jobs, but the method doesn’t scale well. A big flowchart that’s not already in good order will take a very long time to fix.
Giant flowcharts and similar are often not drawn by humans at all. Instead, the links and required format are defined in a text file that’s rendered by software. Graphviz dot is an open source example. Doesn’t do OCR though: Greensands would have to learn how to program dot, and then manually create the definition. Whilst dot is a much more efficient way of tackling hefty graphs than a drawing tool, it’s still lots of hard work. And the learning curve is Intimidating.
If anyone identifies an OCR package that works well in this application, I shall be grateful!
Dave