>>Not entirely true, a properly made PDF contains text as ASCII.
>
>That depends on how you define "properly made". ;-)
>
>A PDF that was made by full-blown Acrobat may have text, depending on the options set within Acrobat. A PDF created by a printer driver like PrimoPDF or CutePDF likely will not. Printer drivers typically deal with dots on a page.
>
>Either one will work just fine in a web browser or in Acrobat Reader. Which one is "properly made"? ;-) Which one is most likely to convert willingly to Word? Which one are you most likely to have to deal with? <g>

True - I've seen printer outputs which basically achieve justification and horizontal character spacing by sending each character to its coordinates. You sometimes get contiguous words sent to a location, then next word etc, but more often than not you get a positional command followed by a single printable character. Which is probably the reason why a HelloWorld.pdf has 14K and doesn't contain the string "Hello, world!". Though, looking into its hex dump, there are some obviously Forth-like constructs, i.e. some sort of PostScript code.