Thou shalt have no other gods before the ANSI C standard 1500

Brian Inglis

snip

When transforming to PDF, the PostScript code is executed and the result translated into PDF-compatible display lists. For example, loops are unrolled, only the followed branch of a conditional is kept, variables and defs are expanded in situ, etc. because those things aren't representable, and THEN the result is pbutted through an LZ77 compressor and sent into the file. Conversion back to PostScript does not recreate that information, because it has been lost.

PDF does paths, strokes, fills of varying types, etc. like PostScript does (it is, after all, a much-simplified lightly-programmable PostScript), but PostScript does many things that cannot be done in PDF. Think of a PDF distiller as a specialized PostScript interpreter that doesn't rasterize, but instead generates display lists.

If you don't believe me take a real PostScript program, "convert" it to PDF, and then "convert" it back. It will be nothing like what went in, only containing the result of the end.

This is not in conflict with what I said. An image need not be bitmapped.

They certainly can be converted to PostScript, which is Turing-complete. They merely aren't, because that's a lot more work than merely implementing a simple PDF viewer in PostScript.

-- GPG: 1024D-9840105A 7102 808A 7733 C2F3 097B 161B 9222 DAB8 9840 105A