Here's some things I've thought of:
- Problem: You have scanned in letters / wills / obituaries/ photos of headstones, etc and want to transcribe them. Solution: A program that would do an initial OCR of images looking for text. It would create a file of what text it was able to extract. Ideally, it would know when it was doing a bad job and allow you to make corrections.
- Problem: You have a bunch of family history documents on your computer and a GEDCOM file. You want to know which people are mentioned in which files. This would allow you to site sources (esp. if you've forgotten to do it or the GEDCOM your grandpa gave you didn't have them) and to see if you already have some of the information you are looking for. Solution: a program that looks for names from your GEDCOM file in specified folders on your hard drive. It would create an index of names and files they appear in. If you were missing date information, it might flag some documents saying this document has a name similar to your ancestors with a date nearby. Maybe also it could add links to the documents in your GEDCOM file if you asked it to. (This idea is based on my one family history experience where a distant cousin called up and asked for the source of a specific date for an ancestor. I text searched a book my dad had put together and found the source. She had the source, but hadn't realized the information was there.)