Cultural Attributes (long) - Feedback Desired

[Glak]

I was going to email this but by the time I finished writing the forums were working again. Hope it formated correctly.

I saw your post about cultures and then shortly after the forums stopped working.
I'm not a programmer but I once saw this algorithim that seemed really good. It is
for reducing colors in an image using three dimensional volumes. I think that it
would work quite well for cultures as well. Imagine the people as colors in a 16
million color scheme (each slighly different from each other) and imagine the
cultures as colors in an optimized 256 color scheme. People naturally fall into
cultures but their individual characteristics remain. Thus if the population changes
the list of cultures updates automatically. Thus hybrids are handled automatically.
The article is at:

http://www.gamasutra.com/features/visual_arts/061997/a_few_good_colors.htm

It also makes sense to use a color model. For example in a picture of some grass,
they sky, the sun and some clouds you would get a set of cultures. Imagine how the
256 colors would be assigned. The sun would probably only get 2-3, the grass
would get more. The sky would get a few but the clouds (especially where the
edges blend with the sky) would get the most. This is probably how it should be
done if you imagine the picture as a representation of the cultures. To encourage
hybrids you would have new people formed by a simple breeding process. Each
parent has a red, green, and blue chromosome. A mate would be chosen by some
sort of process, weighted towards similar chracteristics, for example an orange
person would be more inclined to breed with a yellow person than a dark blue
person. However forces (government intervention, etc..) could cause different
breeding behaviour. After every cycle of population growth cultures could be
redefined and people would be reallotted to new cultures.

Their are disadvantages to this system: no sense of cultural history, cultures can't
be easily named. However I think that with a little work these could be gotten
around. Say there was a religious war between the greenish gray and the violet.
You could grab the current color from each culture (out of 256) and find every
color (out of 16 million) that would fall with into it. So 16 million/256 colors are
affected. Do this for both sides of the conflict. Now assign a marker of some sort
at each of these colors indicating a hatred with the other side. These markers
remain for a while, at least as long as the hatred continues. To find default cultural
hatred bewtween two cultures count up the markers in each (0-256) and multiply
the two values together. Thus if the two cultures don't change and the hatred
continues you will have a hatred value of 16 million. However what happens if the
cultural boundaries move around? Let's say that violet has broken into two
subcultures. Each might contain half of the markers. Now greenish gray hates them
both but not as intensly, although continued conflict (or peace) could change it. This makes sense. Cultures with shifting borders tend not
to hold hatreds as long, on the other hand Arabs and Jews will always hate each
other. I just came up with this hatred marker idea as I wrote it so it is still a bit rough.

[Mark_Everson]

Glak:
Wow, thats a fun idea. Unfortunately, I think the color thing is waaay beyond the scope of anything we're likely to do in Clash. As the model now stands, I'm worried that just having several hundred cultures running around will be too complicated. I think that modeling yet smaller packets of people would be too much. If the game's Sole focus were doing culture right you might be able to get away with it but we have many things to handle.

On the markers... I think they Should be in there in some way. There'd probably be a limit of a few extreme dislikes or affinities that you'd keep track of for each culture. And as you say sometimes dislike should fade away with time.

[Glak]

They way it was desrcibed made it sound really simple. Also aren't you already keeping track of people in some way? As in who is eating what food and working such and such a work site? If so you would just add a little more information. Here is a quote from the guy:

"First, let's take a close look at a color image. Fresh off the digitizer, a typical image uses 24-bit pixels having 1 byte per color component, as shown in Figure 1. Each pixel is one of 16 million variations of red, green, and blue (256-by-256-by-256). The pixels are laid out in a rectangle, 640-by-480. This gives us a grand total of 307,200 pixels per image, with 3 bytes per pixel or just under a megabyte of data."

Each pixel represents a person, each of the possible 16 million colors indicates a possible location for markers. So if there is a large population I guess it would take longer to process it all. However I doubt that there would be 307,200 seperate population people in the game. The information that would need storing would be a cultural color on each person and the coordinates of markers.

The 256 cultures doesn't seem like too big of a number, especially since many of them will be similar until significant hybridization occurs. 256 is also the limit on the number of cultures. That way the late game doesn't get overrun with hybrids. If you set a limit but didn't redistibute then you would either have to cut out old cultures or prevent new ones from being made. The color model means that the 256 most representatvie cultures are always in existance. You could cut it down to something like 16x16x16 for each person and 16 total cultures in the game at any one time. Here's what the guy had to say about speed and such:

"I can tell you the results in hard numbers. This algorithm executes in six seconds on a MacIIx, converting a 640-by-480 image from 24-bit pixels to 8- bit indexes and yielding a high-quality displayed image. A well-written but brute-force solution takes about 70 seconds and yields a lesser-quality image. It took me about three weeks to write and debug the code resulting in 30 pages (1,500 lines) of assembler and a 9K object file."

So how long would this take on a p200 or something like that? Also it is unlikely that the cultural image would involve 307,200 population units.

[Mark_Everson]

Actually the only time people are kept track of explicity (as in who is where) is when a bunch of them gets up to migrate somewhere. Otherwise they are just aggregated as 10pop farming, 3 extracting resources, 3 producing finished goods, etc. And also the proportion of the total population that is in each culture are kept track of. Things are still fluid... we could change it. This particular approach just doesn't strike me as paying off big in playability or wow-factor for the player, YMMV.

Your suggestion is up here now, why don't we see what other people think of it?

-Mark

[Mark_Everson]

Cultural Attributes (long) - Feedback Desired


Cultural Attributes - In Clash everyone's not plain Vanilla

I think the presence of different cultures is a necessity in a game that tries to get historical flavor right. Here's my take on the concept. (Still... Written up sketchily due to time constraints

Cultures are used to facilitate handling the overall character of the peoples in a society. In the game, a Culture encompasses a host of things like ethnicity, language, religion, social structure …

A few areas where I think cultures are especially important are:
· Driving the breakup of multi-cultural empires with the rise of nationalism
· As representing prerequisiste social conditions for industrialization (e.g. a very tradition-bound society will have trouble making the transition to industrialization)
· Conquered peoples "throwing the bums out"

Note: The cultures in clash are just a pale cartoon of what happens in the real world. Some of the notions I'm including would surely be rejected by sociologists and anthropologists. However I think what's below is a reasonable way to simulate at a game level the effects cultures have had in history.

In general each square of the map can have 3 or so cultures associated with it (limitation is memory and speed of game). Each Culture is described as below. New Cultures can be formed by the amalgamation of cultures. For example the Ptolemaic Egyptian Empire would start as bi-cultural, Greek (ruling) and Egyptian. As time goes on, intermarriage and
emulation of the rulers by the ruled would result in a tri-cultural state: Greek, Egyptian and a Greek/Egyptian hybrid.
Depending on how the government acts the cultural change process can be either advanced or retarded. Given enough time the new hybrid culture might become the dominant one in the civilization.

Cultural Attributes
· General culture, aspects not covered elsewhere: language, social norms,… use a polar plot, or some 2D arbitrary mapping for "distance" between cultures
· Unity - How likely are political sub-groups to unite to face a threat to some of them from outside
· Insularity - How unwilling is culture to accept ideas from outside, will it allow other cultures within its empire or must they
conform, or alternately become slaves... also how assimilable is this culture
· Religion - up to two religions w/ pct of pop involved in each, religion, will modify other aspects
· Aggressiveness - 0= pacifistic, 100 = most warlike
· Tradition - Social Mobility combined with how willing the people are to try out new ideas --
0 = free mobility, very willing to try out new approaches, 100 = rigid castes, unshakable tradition
· Social Status of the economic classes: constant sum divided between classes of
1. Upper Classes (Priestly classes, landed aristocracy, upper bureaucrats, higher level military leaders...)
2. Military
3. Laborers
4. Religious
5. Merchants
6. Minorities / Dependents (Children / Elderly / women in some societies)
7. Slaves

I envision the player's guiding of culture as being a main thrust of the game. Money and resources can be thrown into changing the culture (resulting in very slow change). Also some spice can be added to the game by allowing for larger-than-usual cultural change under certain circumstances (points for cool actions...)

People's Thoughts? Will this add spice to Clash or become a micromanagement nightmare? If this is too complicated, is some middle ground preferred to no cultural information as in Civ2?

Hrafnkell:
I wanted to make one observation on this. I think it would be best if all ‘types’ of culture would be defined on how ‘compatible’ it was to other culture ‘types’. This could f.e. be a general number (maybe ranging from 1-100) were a culture ‘type’ 1 woud be more compatible to the one numbered 10, than the one numbered 100. Thus, some religions are more tolerant to others than some, and some culture ‘types’ , f.e. the Hebrews, are scorned and avioded by most other culture ‘types’. This number would then be used in determining the likelihood or extent of two cultures hybridizing.

Mark:
I think I've already got the basis for your ideas between the distance on the "General Culture" plot and Insularity (especially
the last line), which I have quoted below. I'd already chosen a 2D plot for general culture as preferable since there are many
"levels" of cultural interaction. Say we crudely map cultures on a 2D plot going from 0-100 in each dimesion and culture A is at (0,0). If Culture B is at (0,5) indicating very close kinship in language, mores etc the cultures would probably get along well. if B were at (90, 55) well, it wouldn't be pretty. Religion would also play a part.

However, I think History of the interaction between the cultures is just as important. Just as siblings frequenly fight because of being in the same place much of the time so it should be with cultures. So history of the interaction should be included in the cultural specifications, and changed as appropriate as the game goes by...

· Insularity - How unwilling is culture to accept ideas from outside, will it allow other cultures within its empire or must they
conform, or alternately become slaves... also how assimilable is this culture

Peter Dobrovka:
Races/cultures:
How will the AI treat hybrids of 2 cultures? As a 3rd? I ask because they open a bridge for unity but if treated as another "alien" there will be a different outcome.
How will hybrids hybridize? Will they?

The other points are good but their benefit will turn out later I think, when we see the effects on gameplay.

Mark:
Yes, hybrids of two cultures, as I envision it now, will be treated as a 3rd culture, with its attributes largely taken from one parent, the other, or intermediate between the two. Cultural interactions will generally be smoothed by a common or similar
General Culture (including language) and similarity of the other cultural attributes.

For most of the cultrual attributes listed my first take is that hybrids will tend to serve as bridges between the two cultures as you suggest. Take cultures A and B (with A on top politically) resulting in a hybrid AB culture. For each attribute, the worst (most polarized) result in terms for the AB hybrid would be that it was either completely like A or completely like B. In aggregate however AB should get along better with either A or B than A and B do with each other. The things that could change that would be things i'd consider relistic such as AB establishing / endorsing a "new" religion that either A or B strongly dislikes. Another case would be if AB are fortunate enough to become on top politically somewhere and always end up at war with A for instance. This would sour the AB - A relationship. These cases would be exceptions, however, and I think your feeling about the hybrids becoming a bridge is very reasonable.

Hybrids should IMO hybridize with others. If the AB culture is strong politically, it might conquer C, resulting in an (AB)C hybrid. We would probably need some mechanism to keep AB from immediately hybridizing with A or B. Otherwise there will be a myriad of hybrids between A and B. This is probably right in the real world to some extent, but would be a real Mess in a game. Maybe some kind of time lag after AB is formed before it could hybridize with either "parent" again.

[Mikael]

Mark,
Just one question about the cultural attributes: are they likely to change during a game? Features like aggressiveness or insularity should IMO slightly change as the centuries pass.The mentality of your people should be affected by threats or special events of some sort.

Mikael

[Mark_Everson]

Mikael:

The plan is that cultural attributes should drift somewhat over long periods of time depending on circumstances. The player can also Try to push them in the direction the player prefers. However the latter should be expensive and slow to achieve positive results IMO. Some things should be harder to change than others also. If you'd like to suggest details I'm all ears

[Mikael]

Mark,

Maybe at the start of the game you could decide yourself what cultural attributes your civ is going to have.Obviously you would have to make them balanced or if you want you could let the computer generate them.How many civs do you intend to throw in the game?
Maybe at the end of the game if one "culture" is predominant the other cultures (they'd have to be flexible, though) could start copying it and eventually there would be less and less cultures in the world( western culture nowadays has completely taken over).

Mikael

[Mark_Everson]

Mikael:
There is a thread on the Old main BB called "In the Beginning..." that addresses these issues. The culture at the start will be chosen by the player within the constraints of a point system.

I like your idea about cultures trying to emulate more successful ones.

# of civs is essentially unlimited (right now its up to 500). The only limit is the tradeoff of how good you want the AI for Each civ to be, and how long you are willing to wait...

[Lumpkin]

(Lumpkin=Nick Holzapfel btw)

I'm not sure about the aggressiveness atribute. I thing is perhaps a hangover from Civ. I think in the real world no civ is intrinsically agressive. I think aggressiveness should depend on two factors:

1) The government - feudal kings were expected to wage war while democratic leaders are expected to avoid war.

2) A new culture attribute - "Enlightenment". This represents the growing collective consciousness and understanding of the nation, from a simplistic warlike view of the world, to present-day cynical world-view where human life is valued and we have a welfare system to support the needy etc. The higher this is the more likely the culture is to do the "right thing". It would be increased by the power of the media, the living standards of the populace, education standards and humanitarian disasters like the Holocaust and the world wars. Taking Britain as an example several stages can be identified:

1) Feudal England, kings are expected to wage war against the traditional enemies France and Scotland. Human rights are non-existant.

2) The British Empire is the first major power to abolish slavery in the early nineteenth century (I think) and East African slave ports are bombed by the Royal Navy. On the other hand, the empire continues to expand at a phenomenal rate. Not very nice in hindsight but perfectly righteous in the context of the time.

3) Modern Britain: Although mainly as a consequence of the devastating effects of two world wars, but also because people no longer felt the empire was morally justifiable, Britain's gigantic world empire is abandoned and replaced by the Commonwealth.

Also, patriotism seems to be missing. This would make soldiers better and the country more unified. Maybe National Pride would be more appropriate actually. This would be increased by victories or occasional highler popular leaders (Hitler, Roosevelt (cant remember how its spelt), Elizabeth I). This would have the effect of uniting the country in times of pride, reducing disorder and increasing productivity (The early USSR) while times of humiliation would lead to increased disorder as ppl turn from the government in search of groups who promise to restore national pride (post-WW1 Germany)

[Lumpkin]

Oh yeah, another cultural attribute should be "war tolerance". This would depend strongly on enlightenment factor and would increase with a sense of national pride until the ppl are almost crying out for a war (Build up to WW1). However, after economic and humanitarian hardship caused by a prolonged large-scale war recquiring conscription and military defeats the war tolerance will take a sudden dip. This may take may take decades recover (Chamberlain was cheered by many when he compromised with Hiter, supposedly acheiving "Peace in our Time). However, when the nation itself is considered to be directly at threat war tolerance may actually increase (Hitler's bombing of Britain merely increased resolve). War tolerance means ppl are prepared to put up with great hardship without becoming unhappy. On the other hand, if hardship shows up weakness in the government (as in tsarist Russia), the ppl will lose faith in the government.

Which leads me neatly on to yet more proposed cultural attributes:

Faith in Government: When the economy or the national pride begins to decline ppl beging to lose faith in their leadership and turn to rebellious movements.

Respect for the Law: This is a vital part of a culture and has always played a crucial role in any nation's history. Henry VIII's restoration of law and order in his kingdom meant the War of the Roses finally ended and for the first time in a century the heir suceeded to the throne peacefully.

[Mark_Everson]

Hey Nick/lumpkin:

Yeah, I think you're right on aggressiveness.

Patriotism, and war tolerance, I think of as more of a manifestation of government. Are you up to date on the latest govt. stuff?

Respect for Law is already embedded in the govt "class power" structure.

All these are up for debate, that's just the way they are now...

-Mark

[Mikael]

I think lumpkin's "enlightenment" idea is a very good one.It would actually represent the level of "civilization" in your civ.That enlightenment factor would rise as your civ gains experience and wisdom(better education essentially).The more "enlighted" civs would often try to avoid war ,and things like slavery or feudalism would slowly disappear, due to public opinion.
However if one civ decides to continue slavery (that's only an exemple), then the more "enlighted" nations would start to despise that civ, and diplomatic relations would then suffer seriously.
Maybe the more enlighted your civ is, the more flexible and tolerant it would also become, resulting in faster scientific research...
This is very sketchy and rough,but perhaps this idea could be fun...

Mikael

[F_Smith]

I'm ready to put the 'Custom Civ' selection screen together, and I need some definites.

1) please list the attributes ya'll are going to want on the screen -- by name, # of levels, and pt breakdown of those levels. It would fit best if we can limit it to 8 attributes, for now. We can simply do multiple 'Panels', if it runs more than that . . .

Tonight I'll put up the demo as it exists so far. I've got to go to Chicago on Monday, so Sunday night I'd like some at least semi-final answers, at least something to put in the demo!

[Mark_Everson]

I need a little clarification of what you need. Is it the game start screen(s) that you're looking for? If so here are my suggestions

1. World selection (a few std sizes, or go to customize menu)
2a. player civ size [2a,2b already in v0.05]
2b. average size of AI civs
3. position (flank power, or surrounded by other civs. perhaps could be done as # of neighbor civs)
4. starting culture (some general options like militaristic, nomadic, trade-oriented, fuedalistic, bureaucratic, theocratic. if player wants can go to another more detailed screen)
5. Overall tech level
6. Player civ tech level(my thought is that the specific techs should flow from the culture and the general tech level, or could go to a special tech selection screen)

If we do the point cost thing Selections might gray out or something when you can no longer select them. So if you picked few neighbors and a larger civ size you might have tech levels above the overall tech level grayed out because you can no longer afford it. Thoughts...

Do we need specifics for Culture screen etc yet?

-Mark