Re: [Freeciv-Dev] Serious Suggestions

[Stephen Hodge <stephenh@xxxxxxxxxx>]

Matt Busigin wrote:
> On Tue, 23 Feb 1999, Greg Wooledge wrote:
> 
> #I think that if we're going to go to the trouble of adding some sort
> #of extension capability, it should be a full language, not just a
> #simple hack.
> 
> What are the merits of having a full language?  A simple language != a
> simple hack.  It can be a nice simple, relatively powerful and *SPECIFIC*
> language.

With a full language we could extend it more easily to allow complex
tasks to be scripted (eg. a client AI written entirely in the scripting
language - I think this would be really cool). Extending a FreeCiv
specific language originally designed for city build orders is likely to
be more difficult. Of course the simple language is probably easier to
implement. The other advantage with something like Guile is that it is
widely known (and in Guile's case a GNU standard).

> #Someone mentioned "bloat".  I think this is misleading -- we've already
> #got a graphical interface in the client, so adding a scripting language
> #is relatively insignificant.  (I don't have statistics available, but my
> #feeling is that the overhead added by adding a Guile or Python interpreter
> #would be much smaller than the current client memory/CPU usage.)
> 
> OK - I admit that I havn't done any Guile, nor Python embedded programming,
> but I have done embedded Perl.
> 
> It would be much cleaner just to impliment our own simple, relatively
> powerful and *SPECIFIC* language.

Does anyone know what (if any) stuff is available for making a (for
example) Guile implementation easier to add?

Regards,
Steve Hodge