Okay, that's a really lame argument. A graphic is just as much a "document" as a text file. Any distinction there has nothing to do with what PDF stands for. More to the point, they are both equally "like" a PDF--depending on what kind of PDF it is.
In the same way, there is no such thing as a Portable Document Format (PDF) graphic file. What you have is a PDF format file that [b]contains[/b] a graphic image--it's a wrapper around the graphic. Just like the Word file is a wrapper around the graphic.
Not really. You're implying that the text in a PDF file is somehow more "real" or "native" than the graphics in it, like text in Word is more native than an embedded graphic. That's wrong. Originally derived from PostScript, PDF is more like computer code that can describe anything on a page--text, graphics, layout, fonts, and so on. It doesn't favor text over graphics. It's just code that enables an interpreter to re-create the page.
I don't understand why this is so hard to grasp. You get a piece of paper--a contract, a magazine page, a birth certificate--and you scan it. If you save it as PDF, you get a PDF file with code that describes the graphic. If you have OCR turned on, it also extracts the text so you can search it; if not, you just get a PDF file of the graphic, aka a PDF graphic file.
And I don't get your reference to "computer gibberish." Did I say you got computer gibberish?
I'm guessing he may be referring to the "convert to text file" option most OCR apps have. As you probably know when they come across a character they can't process, say the "t" in "cat", they print a series of apparently random characters, resulting in "ca$%gvbLo@*"
One other point. OCR does not always convert to text, for instance products used to read handwritten items like checks or envelopes.