PDF still provides a deterministic scaffold: There are pages, there are objects on the pages, and a renderer can process these with predictable resource use (incl. time). Some of them might then go on and run a VM to do whatever with them, but at worst this means that individual elements will be incomplete.
Postscript is versatile enough that some Mandelbrot set routine gets to calculates how many pages there will be. Eventually. Maybe.
I wonder if someone already tried to embed a pdf reader in a pdf. Would it be able to read itself ?
Wasn’t the purpose of PDF to restrict the turing-complete computation that was available to Postscript? What happened to that goal?
PDF still provides a deterministic scaffold: There are pages, there are objects on the pages, and a renderer can process these with predictable resource use (incl. time). Some of them might then go on and run a VM to do whatever with them, but at worst this means that individual elements will be incomplete.
Postscript is versatile enough that some Mandelbrot set routine gets to calculates how many pages there will be. Eventually. Maybe.
Ah I see! So in a PDF, we /always/ get to know how many pages it has, and not so in PS.
Thanks.
[Comment removed by author]