RFC Europe: Dynamic civ names

[sedna17]

The current model for Dynamic names is somewhat odd. I will have to change some of it so that we will fit it well with the "C++ implemented Python Controlled" scheme. Right now, dynamic names are all in C++.

That's a good idea if it's not hard. There's a bunch of stuff hard-coded in that section of the C++.

If you take a look at the SVN, you'll see I edited the C++ in the renaming code a bit and compiled a new dll, just to test out the XML as I was entering the dynamic names and make sure everything was working. But I support moving that decision tree into Python.

 

[3Miro]

That's a good idea if it's not hard. There's a bunch of stuff hard-coded in that section of the C++.

If you take a look at the SVN, you'll see I edited the C++ in the renaming code a bit and compiled a new dll, just to test out the XML as I was entering the dynamic names and make sure everything was working. But I support moving that decision tree into Python.

There is the process name function that gets called from different events (switch religion/government/year...) My idea was to create a link-list of conditions for every nation and it would be loaded from Python. Then the name module would process the link list of the form: France + Monarchy + Christianity + before year X: Frank Empire. France + Republic: French republic. I believe it is the most flexible setting without creating the huge arrays of names for every civ under all possible circumstances.

 

[jessiecat]

Dynamic names would be great. As would 2nd. LH's and respawnings. As micbic has outlined in his very comprehensive list. Any chance of at least starting to implement them, please.

 

[sedna17]

There is the process name function that gets called from different events (switch religion/government/year...) My idea was to create a link-list of conditions for every nation and it would be loaded from Python. Then the name module would process the link list of the form: France + Monarchy + Christianity + before year X: Frank Empire. France + Republic: French republic. I believe it is the most flexible setting without creating the huge arrays of names for every civ under all possible circumstances.

The most flexible code would allow a combination of two processes. The first is a generic process for renaming civs (which may be what you are describing above). This could involve sticking together some strings so that you get generic names e.g. 'French' + 'Republic' and 'German' + 'Republic' for republics and 'Exarchate of' + 'Vienna' and 'Exarchate of'+'Spain' for vassals of Byzantium.

The complication to this is that we should allow special cases as well. So if a civ doesn't have an XML entry for <CIV_Republic_Name> then we use the generic name above, but if it does, we use the special name. This way, for instance, England with a Republic can be called the "Commonwealth of England" instead of the 'English Republic'.

 

[3Miro]

I was thinking something along the lines of:

struct rule{
int iReqCivic, iReqRel, iReqTurn;
string name;
rule *next;
};

Then in the renaming module:

if ( hasCivic( p -> iReqCivic ) && gameTurn > p -> iReqTurn .... ){
setName( p -> name );
};

Add the details. The rule list would be set via Python.

 

[sedna17]

I was thinking something along the lines of:

struct rule{
int iReqCivic, iReqRel, iReqTurn;
string name;
rule *next;
};

Then in the renaming module:

if ( hasCivic( p -> iReqCivic ) && gameTurn > p -> iReqTurn .... ){
setName( p -> name );
};

Add the details. The rule list would be set via Python.

I'm afraid I don't quite understand this code, but the idea of applying Python-written rules is good, so I approve. "Currently" (i.e. RFC and Panopticon's list) the renaming is done on Civic, Turn (or age), Religion, Vassal status, and number of cities, so these would be the useful variables to have in a rule struct.