I wrote some of the ideas that came to my mind at that moment but suggestions are welcome:
* Support for undo/redo
* Support for a defined language for documents that also configures the used OCR engine
* Implement reading order support
* Improve the generation of ODT files (currently the text is put as text boxes in the generated docs, maybe an option to set it as actual paragraphs would be better)
* Create a plugins system to process the results
* Improve the contents detection algorithm
* Create an assistant-like user mode (like 3 steps to scan and generate a document)
I don’t know if the apps for GNOME’s GSoC should be official modules, something that OCRFeeder is not yet but I think it is an interesting project for GSoC nonetheless.
As for mentoring it, I’m taking some weeks off in June and August but maybe I could still do it.
Let’s see how it turns out…