Re: PDF

[Darrin Smart]

Valeriy E. Ushakov wrote:
> 
> And a rudimentary PDF backend without all the bells and whistles can
> share a lot of code with PS backend.  I think 99.9% of engine code can
> be shared as the imaging model is the same, the differences are in the
> output format.
> 

It would be a little more complex than this. A lot of the PostScript
specifics are actually in the package definitions. You also need to grep
for @BackEnd in these. And actual postscript functions are defined in
some of these too. In addition, if PDF is a binary format it could be
rather difficult to write packages to output them.

I discovered all this writing a backend for OpenPage last year
(http://www.research.canon.com.au/openpage/).

An alternative might be to use the current PS output and use a PS-PDF
converter (Ghostscript?). If someone was to write/patch lout it might be
possible to embed xref-type information in comments and have the
converter understand them.

Or you could write a XML-based specification for the backend which used
a declaritive drawing model rather than a programming language which
would make writing filters to various formats easier.

 - Darrin