Demo 9 Overall Approach

[Mark_Everson]

We have started talking about the Demo 8.1 Goals and Plan in another thread, and the lack of a long-term planning thread became immediately apparent. So let's have all the long-term development discussion here, and reserve for the other short-term thread to more tactical next step discussions.

As an outline of a plan becomes clearer, I will update this post to show what the current state of play is. Let the proposals begin!

[Mark_Everson]

This space saved for future use

[alms66]

The road to random map games...

This is where Clash has been headed for a long time, and we are close to being capable of pulling it off. When you really think about it, there are only three major elements that are keeping us from doing it right now:

1. Save function
(this is essential for those times when something comes up during a play test, or to send in the turn before a reproducible bug occurs, as well as to save the sanity of those who play test for clash and meet one of the projects longest running requests)

2. Reworked Drawing Code
(this is arguably the most important of the three major elements – the reason being that the current code can barely display a 2000 tile map without bogging down due to the vast number of tiles it has to draw – and most aren't even seen on the screen! I don't know what size world map we're shooting for, but I personally would like to see something around the 20,000-tile mark or more and the current code can't even do anything close to that. However, with all of that said, it is possible to do random map games without displaying an entire planet – we could have various types of map generators such as “Mediterranean”, “Desert”, “Highlands”, “Islands”, etc. that generate small, specialized, "local" maps in much the same way that RTS games do, and something along those lines could be our first shot at a random map game, thus making this a minor element to implement, but a very important one.)

3. Diplomacy
(this is also essential to random map games, because without it, you're fighting a never-ending and quickly-tiring war - thus play testers will/do need this to preserve their sanity and to keep them play testing Clash)


There are a number of minor elements which can be avoided for the first try, but should be high on the list following that. These include:


1. Help System
(random map games will likely draw a slightly larger testing audience than previous versions of clash [speculation, but still likely] and a help system [civilopedia] will be needed to help people with various concepts throughout the game to avoid having team members answer many easily answered questions with a simple help file)

2. Setup Screen
(this will be needed to give the player the ability to choose what options to use, what map generator to use, etc.)

3. Install & Patch Procedure [basically along the lines of the testbed currently used]
(Once random maps are a reality, it will be time to drop the 'jar file distribution' approach and do a true installation of Clash – the reason being to facilitate the easier development & distribution of scenarios. The scenario designers could be releasing new scenarios every few weeks and updates to existing ones just as quickly, while the code progresses at whatever pace because – new code simply gets patched in as with all other software and prior to releasing the patch publicly, the scenario designers will have brought their scenarios up to date with the patch so that the new versions of the scenarios ship with the patch as well.)

4. Scenario Editor
(an integrated editor will allow more rapid development and deployment of scenarios)



I left out the 'very minor' details that are pretty much self-explanatory such as the need for random civ placement and such. I think this, along with the list from the 8.1 proposal thread should be the longterm plan for Clash. We should continue to release incrementally (8.2, 8.3, etc.) until we reach 9.0, at which time we should have as much of this done as is possible. “Random maps by version 9” should be our mantra.


Comments?

[LDiCesare]

I agree that Save and diplomacy are the most important.
I'm not sure whether it's the graphics engine or the econ or the ai which will bog down bigger maps, but bigger maps are indeed needed.
Gary wanted to do many of these things: Save + scenario editor and polygons (which would adress the drawing part).

I agree on Help and Setup screen too.

Install and patch I never thought of. I think make installing scenarios simpler would be welcome indeed as you don't want to mess inside the jar when updating the code. Seeing what a mess I had to do to install the Ages of Man CtP2 mod, making things simple for players looks like a good idea.

[alms66]

Originally posted by LDiCesare
I'm not sure whether it's the graphics engine or the econ or the ai which will bog down bigger maps, but bigger maps are indeed needed.

If you generate a 50x50 random map, for instance, and you try to select one of the tiles, the detail map switches a little slow to the newly selected tile, but the tile selector takes 3-5 times as long to redraw. Sure, were only talking 5-10 seconds (at this map size at least), but games need to be fast - no one wants to wait for the computer while they're playing a game. Since there is no population, units or civs on these random maps, I assume there is very little econ and ai code being executed, which leaves me with the drawing code. Now, in addition, econ and ai tweaks will most likely be needed as well, as evidenced by the Ancient Med. Scenario (If you remove the natives, that scenario runs much, much faster making me think it's mostly the ai of the natives - who can see the entire map - taking all the time - in fact the last version of that scenario I made, did just that and should have been released instead of the current version since it was actually playable).

Originally posted by LDiCesare
Gary wanted to do many of these things: Save + scenario editor and polygons (which would adress the drawing part).

Well, unfortunately, Gary isn't around much anymore, so I think we should move on as if he was never coming back. If he does, it will be a pleasant surprise that way.
As for polygons :shudders:, the only way I would ever support the use of something which requires a continuous coordinate system (CCS) is if we went to a 3d sphere, which I don't think will happen, so I don't ever want to see polygons in Clash. Even if we did go to a "globe" representation of the world, I'd prefer to see some sort of tiled globe (and thus not a true sphere) which wouldn't require the use of a CCS.

[Lord God Jinnai]

3d sphere is imo the best, but probably not ideal for demo9 unless we are redoing the map graphics engine. Thus i'd hold off on the engine until later.