Re: [Gimp-print-devel] The requested intro message:)

[Ralph C. <ra...@in...>]

Hi,

> I think for now we should create a single Ghostscript driver, because
> most of the files are shared.  This probably means reworking it so
> that rather than using numbers for everything (particularly model
> names), we use strings (e. g. -dModel=esp870).  I hate this selection
> at compile time nonsense.  In the longer term, Mike Sweet has other
> ideas about how to proceed.

Alongside more documentation about the `good choice' combinations of
GhostScript options for different printers it might be an idea to emit
a warning if an obviously bad set are choosen, and an error for invalid
combinations (this already happens?).

Perhaps this could all be driven from a single table with model on one
axis and the -dDither, etc., on the other.  A rating of 0-5 for
-dDither=2 with -dModel=11 would allow the user to examine the table
and see what he's got to play with, and it could be used at runtime (or
source built from it) to warn the user if say the score was <= 2.  A
score of zero means not possible and gives an error.

I'm all for letting the user do the unlikely even if it doesn't seem
wise -- give them enough rope to hang themselves -- but a bit more
guidance might help them a lot.


Ralph.