We can start by making a data structure that can hold the data for each fleet.

We should begin with just three ship classes (and graphics) and no graphics for the equipment (except for weapons-fire). It won't be hard to come up with some numbers for that small number of items.

Also the AI needs to start simple. If we initially only have two players the fleets could just start flying toward each-other and the AI should just shoot the closest ship.

Here would be the parts:

• Game client
  • Battle sim (place ships on the map where each player would get points to spend on "buying" ships from their loadout and battle to the death!)
    • Graphics for three ships
    • Graphics for weapons-fire
    • Graphics for ship death
    • A rudimentary AI
    • A system of keeping track of each ship's hit-points
    • A UI for selecting/placing the ships for battle
  • Fleet creator (just for making all the different ships in your fleet kind of like a "Fleet Loadout")
    • Data Structure for ships/fleets
      • Ship class
      • Ship class "cost"
      • Ship equipment
        • Which hard-point the equipment is attached to
        • What kind of equipment it is
        • What stats the component has
        • What "price" the component has
      • A UI for creating the ships
      • A UI for making squadrons
• Fleet file exchange website
  • Upload fleets
  • Download fleets

Did I miss anything? (probably a lot in the specifics) Let's get this list of what to do completed before moving on.

Where do we want to start? I suggest beginning with the non-campaign single-player mode because it would allow us to test and develop all the systems needed in the other modes. (except for p2p communication for multiplayer and campaign quirks)

I see 2 main approaches of getting to this point:

Approach 1) We start with getting fleet design and AI working. Once thats done, we can have autonomous fleet vs fleet battles (to battle someone, you just need their fleet file), and let people upload fleet files to tournament servers and we can use the data from what people over use in those to auto-balance the pricing and strengths of components. As soon as real fleet vs. fleet stuff is working though, we can start making (and accepting from the community) various single player maps and campaigns (and potentially mods to the available components and such as well). This pretty closely matches what we would need to drop into the FG MMO to add space combat

Approach 2) Screw fleet and AI design, and just do ship design. Then make a multiplayer 1st and 3rd person space shooter with custom ships (Initially it can not use custom ships). Add in AI to allow for single player then add in control of a full fleet with semi autonomous AI later. (With a bit of work on an modified physics engine and control AI, you could also make an in atmosphere fighter dogfighting setup if you wanted)

Both seem pretty good to me. Approach 2 basically starts as a generic multiplayer space shooter game, then gets features (custom ships, single player, fleet combat etc) where approach one really starts as more of a design and simulation type project. The end result would be the same as once it has all the features from both, we can enable all the play modes: autonomous fleet battle, rts, fps, first person swarm shooter (fps with a fleet instead of 1 ship). All of these could be both multiplayer and single player running on the same engine (if we design it correctly!), and single player campaigns could mix all the play styles as desired.

just single player and player to player multiplayer

Ahh, like in Civilization IV, I see what you are getting at. I see three different modes: Campaign, Multiplayer, and a randomly generated game (like a regular game of Civ with a randomly generated universe). This will make development much easier.

Where do we want to start? I suggest beginning with the non-campaign single-player mode because it would allow us to test and develop all the systems needed in the other modes. (except for p2p communication for multiplayer and campaign quirks)

Let's make a good plan before we jump into development this time. We can make this a semi-fresh start.

Any thoughts?

With my shift to free and open source development tools, as long as we can manage to crawl along with unpaid developers, our costs will stay at basically 0$. Its true that I'm paying a bit for the web hosting, but I use it for several other things too, and its darn cheap.

I agree that having alliances greatly adds to a game. Fortunately for us, once we have things working, adding co-op, and 2v2 etc should be pretty easy. And for the mmo aspect, alliances will be crucial for good gameplay.

If this new project goes forward, there will need to be at least one single player campaign,

Yes. I see this quite broadly, as the single-player campaign could be the main part of the game, or an introduction to the multiplayer aspects like being a tutorial. Games which have implemented fleet-design along with campaigns have been Battleships Forever, Gate 88, Star Knights, Gratuitous Space Battles, and Celetania… I liked their approaches to fleet battle, yet it was really disheartening to see Celetania go bankrupt. Games which require high processing-power on the server-end require investors or high-quality personal servers. I was just dismayed that Celetania never had a success when it hit the market, and I think that this was because players needed the incentive of being on a team to donate, instead of doing a free-for-all. Me on a tangent: The point of donating for non-subscriber games is to obtain power in-game, but if nobody is in your alliance then there is nobody to show-off to, and nobody to support. In the games I have played, people (including myself) have donated primarily to help out their alliances.

So somehow, the concept of alliances should be kept, even if the focus is on 1v1 battles. For example, I am participating in a 2v2 team-elimination tournament tomorrow in a 1v1 cardgame named Alteil :)

I always liked your writing Matt. Great to still have you!

Agreed!

I have some ideas on how to design the user's perspective of space gameplay. We should talk about it sometime.

Hey Tristan :)

The fuzzy focus of the project has also been my fault too. I did not produce much in terms of GC design in the span that it was my responsibility.

I am very sorry about how little time I have spent on FG. Unfortunately college has been taking up all my time. I am trying to structure my time this term so I can actually dedicate some time to FG.

I always liked your writing Matt. Great to still have you!

I have some ideas on how to design the user's perspective of space gameplay. We should talk about it sometime.

As for your new idea, I think its cool. One thing that popped into my head was if you had all these people creating custom AI and then testing them in battles, you could combine all that data and effectively develop a sophisticated computer-driven AI modal. Essentially you could monitor which player-set AI settings produced the best battlefield results, and use that to create computer-bot controlled fleets that are intelligent enough to be challenging in battle. Could be interesting, and would make the single-player campaign more rewarding if you integrated it into that. Just a thought.

Oh yes, defiantly. I have had success with evolutionary algorithms before, and I would love to employ them both for tuning and balancing the game as well as generating optimized fleets. I specifically intend to use tournaments to exploit player's strategic designing ability in helping with this. The idea here is design a fleet, and let it fight, so battles work with no player intervention (its a design contest, though other play modes could be added). At least in this basic mode, it is very easy to run massive amounts of battles with player made and evolved fleets.

For modes with more, evolving AI to combat the player in realtime would be an interesting challenge. I hadn't really though about how far we could take realtime interaction. As it is a short term battle game, fast paced semi first person control is not out of the question even. Interesting. We may have many modes. Really I just was thinking of doing the basic stand back and watch your fleet fight game, but with a system for making sub-fleets obey commanding fleets hierarchically, there is no reason we can't simply swap the highest level command AI with the player and have a multiplayer first person swarm shooter with custom designed fleets. I see a lot of games that could be made on top of a general purpose fleet/ship design and combat engine. (And a single player campaigns using all the different kinds of games would be cool)

FG is really only still around at all because you have continued to work on it. For that reason alone, you have the complete right to take the development in this game in any direction you want. I feel guilty that I haven't spent much time at all on this project in the past year or so, but at the same time I know I've been working on other projects that have allowed me to improve greatly as a coder, much like how you view FG as a conduit for your own self-improvement.

Note also that it is only around because I choose to call some of my projects part of FG and change the design to include them (Planets got added after I decided planet were cool for example)

I'm glad you are hanging around. It's always nice to have more minds working on design! If this new project goes forward, there will need to be at least one single player campaign, so when the time comes I may want to seek out our lead writer for that ;) We can of course have lots of campaigns if desired (they should be really really easy to make)

As for actual progress on the idea, I made some. I decided to implement the library of designs as a tree, a single tree. All your folder/file like organization of saved designs as nodes, fleet designs as nodes with sub-fleets as sub nodes… Tree structure is like this:
Thing - What thing can contain/have as children in tree
Library - Anything
Formation - Formation, Ship
Ship - Hull (only one, Hull subclassed from Module)
Module - Module, Part
Part - Nothing (Parts only specify stats)

I've got a crude tree display working with basic browsing filters (I'm very proud of my filters system, though making a GUI for it will be hard) and color coding, but no GUI to really do anything with it yet.

I'm really starting to like the versatility of this idea. There are a lot of potential game designs that it could work for (heck, you could do a first person ship combat RPG where you fight in tournaments to win money yo upgrade your ship.) The basic design based AI control version to auto balance the game helps a lot too.

FG is really only still around at all because you have continued to work on it. For that reason alone, you have the complete right to take the development in this game in any direction you want. I feel guilty that I haven't spent much time at all on this project in the past year or so, but at the same time I know I've been working on other projects that have allowed me to improve greatly as a coder, much like how you view FG as a conduit for your own self-improvement.

As a college student who is now expanding my interests and beginning to found businesses and projects of my own, I know that I won't be able to devote any time to FG from a coding perspective. However, the conversations I have had on some of these forum threads and the debates over design theories we have engaged in have greatly improved my own thinking process over the years and have been honestly fun. Because of that, I'd like to say that I'm always open to bounce ideas off of or to just talk to about anything. I still follow FG's RSS feeds, so as long as stuff is being written on these sites I will be listening.

Although I probably won't code at all, I love to write and so in addition to posting on the forums I would also be willing to continue developing the storyline or writing descriptions for the game if its ever needed.

So yeah thats where I stand on taking FG in a new development direction.

As for your new idea, I think its cool. One thing that popped into my head was if you had all these people creating custom AI and then testing them in battles, you could combine all that data and effectively develop a sophisticated computer-driven AI modal. Essentially you could monitor which player-set AI settings produced the best battlefield results, and use that to create computer-bot controlled fleets that are intelligent enough to be challenging in battle. Could be interesting, and would make the single-player campaign more rewarding if you integrated it into that. Just a thought.

Fragmented Galaxy has been an extraordinarily poorly focused and organized project. I know this, and it is mostly my fault. Also, I don't really mind. For me it has always been an impulse based project. I felt doing ground combat first would allow us to quickly get a functional game out. Without ground combat, there would be no resources, and thus no economy, ships, or game. Well, I went and made planets for a while instead.

Anyway, the plan for ground combat, which will be the source of all resources and ship production, really hasn't been designed to a point where I feel I can implement it. The designs just seem to gain complexity, and I'm unsure of how big of portion of the game screwing with ground combat stuff should be. Originally the plan was to make is super simple, but the idea of significance between different parts of planets was introduced, and things got messy. On that front the project has basically stalled.

Now, on a different topic: Why do I work on this project? The concept for Fragmented Galaxy (aside from the unrefined ground combat ideas) seems very elegant from my perspective as a programmer. It has a very simple yet powerful and realistic economic system, very powerful political system based on some simple building blocks, a rather different and clever technology/research approach, an truly clever and powerful AI setup integrated with the fleet design which mirrors the ship design system. All of these are built on simple principles that are combine either through a hierarchical structure, networking or both. From the programming perspective Fragmented Galaxy is in some cases a worst case game (massively server side, computational expensive, realtime), but those are issues related to computational load. As far as basic implementation and design, Fragmented Galaxy is just a collection of fundamentally beautiful algorithms. This is why I'm still here, 3 years into the project.

And, as to why the project always gets sidetracked: I'm just here to play with clever algorithms. When I started generating planets, I ventured into the realm of procedural terrain generation, high end rendering, graphics card programming, and now the Python programming language. Any of these areas could happily consume years of my life, and I expect them to do so. As far as Fragmented Galaxy goes though, none of these are really very important to the game.

And that leads me back to a new idea. I'm going to want to get to play with those clever AI, ship and fleet design ideas sometime, and ground combat just isn't seeming very interesting at the moment. Thus, I came up with the idea of making a fleet battle game. Basically you would be given a fixed budget to make a fleet, and battle another fleet made with that budget (or potentially a different budget as a handicap). This could start without fleets, just with single ships, and fleets could be added later. Saving fleet/ship designs (including AI settings) out to files for sharing would allow a single player campaign (which could serve as a tutorial) of preset opponent fleets and corresponding budgets (you could score based on budget used, or losses taken, or both, or simply win/lose). We could allow players to battle their own fleets, battle over the network (Lan) or online, and we could host official tournaments (submit your fleet, battles run server side)

This would be a pure strategy and design game with no realtime aspects. Simply make your fleet (or select one you already made), and press go. Potentially realtime or turn based modes could be added where you could send in backup waves or something like that.

This entire system could be kept intact, but imported into Fragmented Galaxy to allow you to use your same fleets in the MMORTS. The only issue I see is with techs; not all nations in Fragmented Galaxy will have the same techs, and none will have the exact same development status on the techs. It is only an issue with integrating the fleet battling game with FG, and thus I'm not going to worry about it now.

Anyway, I now have an idea that will let me play with the AI, ship and fleet stuff, and make an interesting game that does not have serious hosting issues. It might happen sometime.