Add a "Core" nation set of 50 playable nations.
If players only pick nations from this set (or none at all), then other
nations in the game will only be picked from this set, otherwise the
full range of nations will be used.
Nations were mostly selected for the "Core" group by appearance in
similar Civ games.

Original patch by J.M. Maalderink (mixcoatl@gna) with changes by me.

See gna patch #3432.

(Browse SVN revision 21759)

As all nations still live in same .po file, maybe you should add at TRANS comment next to each core nation legend so translators could search for those.

> Is it possible to have 'extended' civil war nations spawn in a
> core nations game?

2.4 (patch #3449): not unless all other "core" nations run out, so, mostly not.
2.5 (patch #3448): not at all.

> If not then maybe it would be a good solution to add some
> non-core nations as non-playable nations to the core set,
> giving every core nation at least one related civil war nation.

Hm. I'm not sure the current solution would permit that. presumably you'd want these "civil war" nations playable when using extended nations; one goal has been to avoid having to fork any of the individual nation ruleset files; but having nations playable in one context but not another would require that, I think, unless we invent even more syntax. Plus I'm not sure that non-playable-but-civil-war-able is supported currently (haven't checked).

Having extra unplayable "civil war" nations in core would meet the goal of not overwhelming players, but would compromise the one about not overloading translators -- they'd have to translate nearly double the number of nations. (I suppose you could ignore the nation legends for these, but any po-file-splitting arrangements are unlikely to make that distinction.)

> It seems I missed the Spanish... they should be in of course.

Who gets the chop from the "extras" to make room?

>> [scenarios]
> Gallic could be represented by Celtic, British by English and
> Cambodian by Khmer. [...]

And I forgot that modern scenario format allows multiple nations to be candidates for a starting position, so we can keep the existing nation/position mappings as well (may require updating the format of scenarios).

(BTW, another use for this core group would be if a custom tileset wanted a custom flagset -- IIRC one tileset author on the forum wanted their own flagset for some reason. This allows them to focus on a manageable number of flags.)

Patch #3449 is now committed.

I'd like to get this in for beta1, if possible. Are we happy with it the way it is, given the status wrt scenarios/civilwar/conflicts already posted?

I'm happy to commit what we have to trunk and S2_4 myself -- I'll do it before beta1 if no-one says anything. (But I'm equally happy for someone else to pick it up.)

Some analysis of conflicts and civil war nations within the proposed set.

Civil war nations: 32/50 of the core group have no civil war nations within the core group, so after my proposed code changes, they would spawn a random probably-unrelated nation. (Many of the other 18 have >1 civil war nation within the set.)

Conflicts within the set: 11 nations have conflicts:

• Byzantine: Greek, Roman
• Danish: Viking
• French: Roman
• Greek: Byzantine
• Hunnic: Russian
• Italian: Roman
• Portuguese: Roman
• Roman: Byzantine, Italian, French, Portuguese
• Russian: Hunnic
• Swedish: Viking
• Viking: Danish, Swedish

Re my comment #1, I've got some proposals for code changes to avoid these problems:

• Long-term (for 2.5): patch #3448
• Short-term (for 2.4): patch #3449

I assume this is at least partially in reaction to the recent discussion on -i18n.

I had the same idea, and I think something like this is the most we'll be able to do for S2_4. Some drawbacks:

• Even if the single player, or all human players, pick nations from the "Core" group, there's no guarantee that all nations in the game will come from it (although I assume the match=3 tries to encourage this). This becomes a problem if we say it's OK for translators to only translate the "Core" nations -- as a player, there's no way to ensure that you don't suddenly have untranslated strings thrown at you from non-"Core" nations. This could presumably happen if matching falls back, or perhaps more likely, in case of civil war.
  • I've been wondering about doing something more radical on trunk: having a nationset choice that is orthogonal to or a subset of ruleset choice, so instead of selecting from group "Core", there's a separate drop-down where you select "Core" or "Extended" nations, each with their own groups. That way you guarantee no nations outside the set will ever be chosen. I haven't thought hard about what to do about civilwar nations pointing outside the set though, and I doubt it can be done for 2.4.
• Conversely, the match=3 will presumably distort random nation selection where the user is happy with using the extended set, such that the same old "core" nations will tend to come up over and over rather than the full richness of the extended nation set.
• "Core" is a bit of a soulless name for a nation group, but it's exactly what I've been referring to this idea as in my head, and I haven't thought of a better suggestion.

> It consists of nations that appear in the games from the
> Civilization series:

We've already got some of that with the separate civ1/civ2 nations.rulesets that appeared in patch #2243. Presumably you list includes Civ3 and up too.

> Celtic will be added to the game in patch #3316

[...]

> Is this still on time for 2_4?

I don't see anything in this patch that can't go on S2_4 in principle, if we accept the drawbacks (although I don't think this patch will apply cleanly, the differences should be trivial).

WRT patch #3316, we said we wouldn't put any more nations in S2_4, but I daresay we could make an exception given the rationale.
(Note that civ2/nations.ruleset already has a "Celtic" nation, this being the only one not covered in the default nations -- it's quite different from patch #3316.)