I'm not sure we should use exactly the same file as civ3. For example, png are easier to use with C#/GDI+ than PCX. But we could use the same structure (suppose we want the same type of map, we can have the same terrain file), and then it's a simple matter of converting the format.
Some for flic.
The question then is "should it be exactly as civIII", or what should we change?
We need to define a realistic list of changes. If it's just to end with the same limitation as civ, then you have your game, it's named civ III.
Before actually coding anything, we need to make a good design specification: what we want to do, what is "necessary", what is overkill, how moddable we want things, etc.
Then from this we can define the main structure, how the engine will work, what strategy for data structure/editor, what type of UI...
If we want a relatively simple map with only a few animation at the same time, then GDI+ can be sufficient. If you want something complex (like dynamic forest made of many independant trees, or units made of several little guys like in CIV or Civ V, then XNA is a lot faster, but it's a lot more complex to code, at least for me).
What is reallistic IMO, supposing we have a motivated team with sufficient time to work on it (not necessarily big, I think 2-3 developpers, a project manager, 1-2 artists, and a few players to contribute design ideas and test, it's sufficient, and some people can have several roles), is to make a kind of clone with similar map quality, similar gameplay, but without an AI (so every civilization is played manullay by a human player in hotseat).
Kind of proof of concept.
Then we could have AI or multiplayer later (but this would not extra skills, and with extra skills it can be done more or less in parallel).
Without an AI/multiplayer, the data structure can be quite simple, the algorithm for movement/battle etc are also quite simple.
And the map engine is not to hard.
There are basic design issues to solve, like:
- Iso square grid, or hexes?
- Map using a notion of altitude, or remains "flat" like in civ?
- Scale of the map:
-- Only one scale for map (with strategical and tactical levels handled there?), as in civ III?
-- A strategical map, but where battles are handled in a different UI, with a minimum combine armed effect, as in Call To Power?
-- One map for strategical, and then a second map for tactical battles like in Heroes of Might and Magic or Total War (without being as detailed of course).
- For units, buildings, etc: are they all independant (each attributes needs to be edited manually for each item), or is there a concept of inheritance? For example, when I started my game, I had a notion of class/subclass/type for the units. It was possible for example to say "attack = 5 for class infantry", then "legion is the same as infantry, but with +1 for attack".
This gives a lot of flexibility, but complexify things a lot (probably too much!)
As you can see, there is a lot of things to decide before coding anything.
Open Source I'm not a strong believer of that at first. I think it requires something less public to get started and have good basis, and then later why not opening it for "expansion".
Some for flic.
The question then is "should it be exactly as civIII", or what should we change?
We need to define a realistic list of changes. If it's just to end with the same limitation as civ, then you have your game, it's named civ III.
Before actually coding anything, we need to make a good design specification: what we want to do, what is "necessary", what is overkill, how moddable we want things, etc.
Then from this we can define the main structure, how the engine will work, what strategy for data structure/editor, what type of UI...
If we want a relatively simple map with only a few animation at the same time, then GDI+ can be sufficient. If you want something complex (like dynamic forest made of many independant trees, or units made of several little guys like in CIV or Civ V, then XNA is a lot faster, but it's a lot more complex to code, at least for me).
What is reallistic IMO, supposing we have a motivated team with sufficient time to work on it (not necessarily big, I think 2-3 developpers, a project manager, 1-2 artists, and a few players to contribute design ideas and test, it's sufficient, and some people can have several roles), is to make a kind of clone with similar map quality, similar gameplay, but without an AI (so every civilization is played manullay by a human player in hotseat).
Kind of proof of concept.
Then we could have AI or multiplayer later (but this would not extra skills, and with extra skills it can be done more or less in parallel).
Without an AI/multiplayer, the data structure can be quite simple, the algorithm for movement/battle etc are also quite simple.
And the map engine is not to hard.
There are basic design issues to solve, like:
- Iso square grid, or hexes?
- Map using a notion of altitude, or remains "flat" like in civ?
- Scale of the map:
-- Only one scale for map (with strategical and tactical levels handled there?), as in civ III?
-- A strategical map, but where battles are handled in a different UI, with a minimum combine armed effect, as in Call To Power?
-- One map for strategical, and then a second map for tactical battles like in Heroes of Might and Magic or Total War (without being as detailed of course).
- For units, buildings, etc: are they all independant (each attributes needs to be edited manually for each item), or is there a concept of inheritance? For example, when I started my game, I had a notion of class/subclass/type for the units. It was possible for example to say "attack = 5 for class infantry", then "legion is the same as infantry, but with +1 for attack".
This gives a lot of flexibility, but complexify things a lot (probably too much!)
As you can see, there is a lot of things to decide before coding anything.
Open Source I'm not a strong believer of that at first. I think it requires something less public to get started and have good basis, and then later why not opening it for "expansion".