Re: Difference between ps and pdf

[Valery Ushakov]

On Wed, Apr 06, 2016 at 09:32:09 +0200, Matej Cepl wrote:

> On 2016-04-06, 02:40 GMT, Valery Ushakov wrote:
> > PDF backend is not really suitable for anything advanced.   
> > It's ok for text and tables, but graphics doesn't really work.  
> > Advanced graphics in Lout relies on PS interpreter to do 
> > computations to "finish" the layout.  This obviously can't 
> > work with PDF. 
> 
> My question is whether Lout generated PostScript travels without 
> problems to PDF via ps2pdf. Could one be certain that whatever 
> tricks they make in PostScript will carry to PDF?

Absolutely.  This is what distillers do by definition.

If you need reassurance - the user manual that uses all lout features
and distills ok :).

You can see obvious examples of this in graphf.etc, e.g.:

    def "*"
        precedence 35
        left x
        right y
    {
        x y "mul"
    }

etc to "compile" infix notation of Lout expressions into postfix PS
code.  Or

    def xloop
        named from { 0 }
        named to { 0 }
        named by { 1 }
        named do named x { "xval" } {}
    {
        from by to "{ /xval exch def" do "} for"
    }

to emit PS loops.

PDF backend prodives only a very simple hack for basic graphics (see
READMEPDF).  And even when PDF backend scripting hack is powerful
enough to what you need, you have to to explicitly write

    @BackEnd @Case {
      PostScript @Yield {
        { ... ps code ... } @Graphic x
      }
      PDF @Yield {
        { ... pseudo code ... } @Graphic x
      }
    }

which is messy and error prone because the code is duplicated.  So
it's much easier to just use PS and distill to PDF.


PS: (no pun intended) Still, the standalone PDF backend, warts and
all, is quite handy when you need it and understand its limitations.

-uwe