Re: Configuration options [was: Policy question - to PM3D or not to PM3D?]

[Petr M. <mi...@ph...>]

>> PM3D support adds image processing but that is an option in v4.0.

PM3D being a configure option dates to its beginning in 1999, when it was 
just an option to draw colour maps and surfaces (well, the "image 
processing" is not the best term for this).

>> Users who have no need for the image processing features, who do
>> not have graphics hardware to support PM3D graphics, or to whose
>> OS PM3D is not ported can build a v4.0 executable without PM3D
>> support.

PM3D features can be used on any platform -- and if even not natively via 
a screen terminal, then via postscript or png.

"Not having need for pm3d" is not a good excuse for not compiling it in. 
E.g. the fitting module is not used by many people and it's there as well, 
etc.

>> You seem to be taking the position that portability and local
>> configurability of gnuplot are no longer goals,

Yes it is, but pm3d is a functionality of gnuplot, it does not depend on 
external tools, thus it does not need system-specific configure.

> plots with portable color assignments, filled curves, color-coded
> scatter plots, and so on.

This shows that the PM3D has propagated into so many places that that's 
really just the time to remove its #idefs.

>> I believe requiring gnuplot users to include PM3D support is a major 
>> change that may affect portability.

Don't you rather mean that NOT including PM3D decreases portability?

> I would be more impressed by that argument if someone could point
> to a single platform for which inclusion of PM3D  support is problematic.
> Are you agreeing with HBB that a 10% size difference is a make-or-break
> issue?  Is there reason to believe that the size difference on non-unix
> platforms will be greater that 10%?  I encourage someone to run the
> same test builds I did on, say, MSWindows.

I don't think that's worth it -- mainly if GNU C compiler is used.

---
PM