Re: Re: Win32: change pointstyle

[<ds...@ch...>]

> 
> From: Petr Mikulik <mi...@ph...>
> Date: 2005/10/07 Fri PM 06:03:11 EDT
> To: Bastian Maerkisch <bma...@we...>
> CC: gnu...@li...
> Subject: Re: Win32: change pointstyle
> 
> > This patch aims to increase consistency between the windows
> > terminal and other terminals:
> > The order of the pointstyles is changed. It is now the same
> > as in gd.trm, pdf.trm, post.trm and possibly others.
> > And default colors are now taken from web_color_rgbs[], as
> > do gd.trm and pdf.trm. Btw. terminals are not very consistent
> > here: gd.trm uses all web colors, pdf only the first 12
> > and postscript uses 9 or so. win.trm uses 15 different colors ...
> 
> There was an effort before 4.0 to unify colors and pointtypes "as much as 
> possible". It would be great if this effort continues. Colors are still not 
> well unified as you've found.

I agree with this.  It doesn't bother me too much because I think the most popular colors of the few terminals I use are about the same.  But it can be annoying to find that a line is yellow when working in a window terminal and brown when when making a hard copy.

Would it be too much to devise some kind of color table lookup for each terminal?  Then have some kind of "get_color(term,..., char *color)"?  E.g.,

lookup_table {
  {"r$ed", 1},
  {"y$ellow", 2},
  {"m$agenta", 3}
...
};

Then, when a user issues a command requiring a color s/he can say "yellow", "green", etc.  Gnuplot then does a

if (!(color_number = get_color(term, user_char_string)) {
  printf("sorry user, don't know that color");
  errorout();
}

That way, all the headache of color matching is left to the person creating the terminal driver.

Of course, we'd have a "show term colors", etc.  There'd be variations on this naturally.  The one down side would be a user utilizing colors for one terminal that are not in another terminal.  Switching to the new terminal would cause faults.

Dan