Re: [Xconq-hackers] A modest proposal to simplify variant development

[<ms...@an...>]

On Fri, 27 Apr 2007, Massimo Campostrini wrote:
> If a unit (or terrain, or material) type cannot appear in a game
> (i.e., it is not present and it can't be created later on), don't
> report the type at all in the interface, neither in the help screens,
> nor in the unit/material summary...

I think the current interface is actually meant to work that way; but as
you've noticed, it doesn't quite work.  Part of the problem is that it's
not necessarily obvious which units and other types can or can't exist.
To get the right answer you have to trace recursively through which units
exist at setup time, which units they can construct (including issues like
"this unit supposedly can create that unit, but doing so would require
more construction points than it can ever actually have"), which units can
change to other types as a result of wrecking and capture, and so on.  I
think the current interface tends to err on the side of caution - it shows
types if it thinks there's ever a possibility of the type existing, even
if it can't actually.

This would be a fun use for Prolog, actually - we could get the whole
theorem-proving thing going for whether a given type can ever exist.

Anyway, I'd certainly be in favour of improving this aspect of the
interface, though my expectation is that it might be tricky.
-- 
Matthew Skala
ms...@an...                    Embrace and defend.
http://ansuz.sooke.bc.ca/