Re: pslatex code contaminating the gnuplot core

[Hans-Bernhard B. <br...@ph...>]

Ethan Merritt wrote:
> On Sunday 13 March 2005 11:34 am, Hans-Bernhard Broeker wrote:

>> The issue is not how they're defined, the issue is which parts of
>> the programs know about, and use them.  The terminal API shouldn't
>> have to know about the meanings of individual layers' designators.

> But that is exactly what the epslatex driver *does* claim to want.

The wish to do that is not entirely unreasonable.  But the method to
fulfill it needs reconsideration.  A proposition:  let's say we do add a
term->changelayer(int layer) function, to be called by the core at
appropriate places (and in strictly ascending order of 'layer').  The
numbers passed to this function would have a meaning only to the core,
but be documented somewhere.  What currently has options 'front' /
'back' would become layers 1 and 30000, e.g., and be documented as such.

The epslatex driver could then output its \includegraphics line as the
layer number crosses a certain number (hardcoded, possibly configurable
by a 'set term' option).

> Is it really so important to be able to write labels behind the plot?
> 
It could be --- that's why we *have* back/front options on 'set label'
and other commands, in the first place.  A more obvious candidate would
be the on-axis tick labels in a plot like this:

	set xtics axis
	set xzeroaxis ; set yzeroaxis
	splot x*y

Having these labels appear in front of the plotted data would be
visually very disturbing, creating an Escher-like appearance.