Got here from here:
http://gsfn.us/t/2mybk
I found out the hard way also, converting a document expecting actual text, not a bitmap image. It was frustrating, actually, to have to figure out the difference. I would suggest at the least, forget the combo drop down box and use radio buttons so that the 2 options are "out there" and add a popup help item to it, something like that.
In reference to my specific issue, there should be an option to "force" an OCR in spite of there being text objects already in the document. A variety of reasons exists why a user would want the OCR engine to parse over a document again, especially if a page is re-scanned or updated. Right now, the logic of the OCR process is very limited, if not "broken." Why did I pay extra for this feature when I cannot use it but once in a document if saved after an OCR run?
Also, again, it would sweet if the optimize function/remove objects included text objects/OCR data. That would strip a document of all previous OCR data. Seems like a very useful function to me. : )