Is this a one dimensional form of OCR? the difference in aka screen 0 printing
Not really. _PRINTWIDTH, from I remember, basically prints your text on a hidden blank screen 10,000 pixels wide and _FONTHEIGHT tall. Then it starts at the 10,000th pixel and looks for a white pixel (text), instead of a black one (empty space). When it first encounters the white pixel, it quits and returns that as a value for the printwidth.
It's a hack that works, but it's slow. (Imagine a font 20 pixels tall, and your text is only 123 pixels in width in total... That's 20 (pixels per vertical line) x 9877 (vertical lines) comparisons before it returns 123 as your solution.
Compared to a monospaced font which is (character count * fontwidth) in size.
No real OCR going on anywhere; just a hunt for color on a blank background, but calculating non-monospaced text width is a slow command, as illustrated via my demo above. If you need speed, you'll want to write your own custom width calculation routines.