Terrain

[alms66]

I still think what is shown on a tile should be a function of a 'View Filter' (Economic View, Political View, Military View, Population View, etc.), with a second filter that works as the 'Map Filter' does, as a qualifier showing a particular variable using color hues.

The current 'Map Filter', "Riots," would then require setting the view filter to 'Population View', and setting the map filter to 'Riots' (I'd use "Rioting Strength" instead, since "Riots" isn't very descriptive, btw).

Further Examples:

[b][u]View Filter     | Map Filter[/u][/b]
Economic View   | Taxes Collected
Economic View   | Production, Food, or Resources Produced
Economic View   | Trade Activity (active trade routes)

Political View  | Diplomatic Relations: (Trustworthiness, etc.)
Political View  | Diplomatic States (Shows diplo states for the selected empire)
Political View  | Active Treaties

Military View   | Power Projection (shows how far and wide, and at what strength power can be projected)
Military View   | Supply

Population View | Riot Potential
Population View | Ethnicity: Group1 (one for each group available - also Religon & Social Classes done the same)

Some of the example Map Filters I have listed above would require 'special case' coding, which would probably put them at the lower end of the priority scale, but many of them are simple queries of a single variable, which makes them much more appealing to code soon. Getting the View Filter on the toolbar, even if it does nothing functionally except change the available map filters based on the selected view filter would be a good thing to do, as it would give a slightly better preview of the overall system than the single Map Filter button and entry we have now.

[Mark_Everson]

Hi Laurent:

The stars look reasonable, and certainly give the function that I wanted, thanks for doing it. The swords look pretty good too. For the cases where they don't show well due to color, perhaps putting a black background around the swords would help.

Hey Alms:

Picking a few of these to implement to flesh out the options as you suggest sounds good. Taxes FE would be easy to do.

[LDiCesare]

I agree with the map filters.
I have a doubt avout trade routes, though, they are routes, so more of an overlay (if you know the way it was shown in Ctp2).
Politicla views I agree. I just checked a demo of a game (Knights of Honor) which has such a feature, it's very useful. They add icons to show alliances and such treaties/possible treaties, which would also be possible.
Power projection: Yes. I've started coding that which is needed for the ai to "see" that. I'll need some thinking in order to show the overlays better. Currently, the code uses a context obejct to decide what should be shown, and there is a MapAI class that holds the data the ai (or player) sees. I will have to make them collaborate better. This class already has an economic richness handle inside, used by the ai to evaluate if targetting a square is worthwhile. That would allow a filter for "gross economic value" of the square.
Riot potential should be ok to display but again, I should probably dump the figures into the MapAI, which is not currently the case, in order not to recompute them everytime the ilter is switched.
Ethnicity would actually need 2 different kinds of filters:
One to show the distribution of a known ethnicity (thus many filters), the other to show majority ethnicity in every square.

[alms66]

Originally posted by LDiCesare
Riot potential should be ok to display but again, I should probably dump the figures into the MapAI, which is not currently the case, in order not to recompute them everytime the ilter is switched.

Since the player has to actively request the information from the map filters, I don't think recomputing every time (each turn, or twice a turn at the most) it's requested is a bad thing. This is not something you'll request every turn most likely, though some people may play with this filter on for the entire game forcing a request every turn.

Taxes, Food/Resources/Production, Ethnicity and riot potential seem like good candidates for implementing. There's the other list somewhere which might offer some possibilities that are just as easy.

[LDiCesare]

I don't think recomputing every time (each turn, or twice a turn at the most) it's requested is a bad thing.

The real problem is I always compute them at least once. It would just be storing one figure in the map, and the ai, if not the player, may want to check it every turn.

[yellowdaddy]

...I'd still like to see a language filter to go with the religion one

{- as they are both constituent parts of "ethnicity", language more so than religion IMO (as discussed at length earlier) because the language is the primary source of the names of each EG, and the primary differentiator, as it's a layer that sits over race and religion.
It may be meaningless to some, but it adds flavour and authenticity IMO, but that's because I'm eager to see a degree of simulation to make the game deeper than all other 4X-style games.}

I know you might say, you'll add it later, but I think if you're planning for this now, it's the best time to put it in.

[alms66]

I believe language and culture in general could play a much greater role in Clash myself. Having a set of values for culture is fine, but there is much more to culture than the few qualities measured currently. I guess you can consider that a vote for languages and expanded cultural modelling in general.

[LDiCesare]

Here's an example of an economics filter.
Two things of note:
1) There's a problem with coastal tiles.
2) The importance of population is huge.
The second point is linked to the fact that econ value is based on population.
The first point is due to the fact that the water extends farther from the diamond. I should probably be able to fix it if I filter before the coasts are added to the image.
(Edit - The white parts are due to a problem in the image grabbing/conversion to gif)

[alms66]

Which Economic filter is this exactly?

[LDiCesare]

This is the "economic value" of the square. The one used by the ai to evaluate the worth of a square. It's 20*estimated tax revenue + production tax base.
I will add the tax revenue filter (without the other part) too.
The filter as is shows huge blue squares in highly populated areas.
I also added a filter for possible riots. The filters may need some refreshing from one turn to the next, and the economic value on first turn is not computed, which is also annoying.

[alms66]

OK, given the fact we're computing economic value, shouldn't the tiles with the least economic value be very light, while the others are very dark? I see several which have no color at all, unless my eyes just can't see it in the screenshot. That seems reasonable for water tiles, but I would expect every land tile to have some color, given the filter.

Secondly, why'd you switch to blue?

[alms66]

bump...

[LDiCesare]

I picked blue because I wanted somehting positive and red rings as danger to me.
The tiles that have no color in fact have some, but the image shows that, actually, most tiles are quite useless in this map (and in fact most maps): Only the tile with a lot of population generates some economic output. This may mean that the maps are not balanced or that the ecobonlics model puts too much emphasis on population.
As far as color is concerned, I can do a red filter, green or blue easily, a grey to (red/green/blue) easily, or any other colour with more work. Whichever looks better to you is ok for me as I can't really judge that well.
Luminosity shouldn't change on a tile, though I may have made an error in the formula, but blue always looks darker than other colors for some reason. I wish java had better api to handle hue/luminosity instead of RGB, but the color api is mostly horrible to deal with in my opinion.

[alms66]

Well, I'd still really like to see the red to blue scale for the map filters, where red represents high levels (whether good or bad) and blue represents low levels, and a mixture of red and blue (some shade of purple) indicates a middle level. In fact, you could just as easily replace the above red and blue with any colors you like, just so that it's always the same. It's simple enough to become second nature to everyone once you get used to it, and it is consistent, rather than each filter using a different color. Not to mention the fact that you could write the filters quicker because you'd be able to just change what variable you're querying for.

[LDiCesare]

The color is just an argument to the constructor, so it's quite easy to change. Blue to red can be done, the opnly problem for me was I equate red with bad (so riots = red).