Difference between OCR Search,Scripted and External

I have found three options as OCR Search,Scripted and External as Test Engine in ‘Generic OCR Platform’ under Preference tab
I know how to use ‘OCR search’. Where I can find help for other two :?:

OCR uses an OCR engine to look at the pixels on the remote screen and figure out what is text and what that text is.

The external TIG uses a Text Image Generator executable on the SUT (or another similarly configured machine) to create images of text using that machine’s native system calls. The images are then returned to eggPlant to be used in searches. These images are cached, but not written to disk.

The scripted TIG is for machines (like Linux) that we haven’t written a TIG for. The idea with a scripted TIG is that you create an eggPlant script that uses an application like a text editor, uses TypeText to type the desired text, and then captures an image of that typed text. Again, the image is cached, but not written to disk.

Our expectation is that the OCR approach will largely negate the need for the other two mechanisms. The major exception would be if you are using a font with characters that are highly stylized and don’t look like “normal” text.

These approaches are described in the Using Eggplant manual and in the Eggplant Reference.