Planet
navi homePPSaboutscreenshotsdownloaddevelopmentforum

Version 7 (modified by nicolasc, 17 years ago) (diff)

Notes from Thinktank IV

The two papers that were available are use as basis with added comments.

Pädu's Proposal

  • Be sure to define small goals in the Thinktank ans to be able to approach them iteratively.
    While this is a good approach for an existing framework, a lot of base coding need to be done. See below
  • Try to define the most important elements needed for a minimalistic running game. Some of them you already achieved. My opinion: Focus on interactivity. Depends on:
    • simple AI (spaceship follows pilot and shoots on it, probability of hitting it is for example ~1/d (d is distance to craft))
      Martin did quite a good Job in creating an expandable base for the AI in a 3 layer approach. Following and shooting are probably the important parts for the pilot, coordination, target selection for the squad commander.
    • simple collision detection (bounding box is sufficient for a first approach, _don't_ code your own collision detection alg.)
      We try to get ODE working, and incorporate it as physics and collision engine.
    • simple mission goals (kill all pirates, kill 5 aliens, etc. just simple counters)
      Full ACK, start simple, extend as the game grows.
    • simple weapon (one is sufficient): sound and vision
      We still need a good weapons system, thus the approach would be, create a weaponsystem, and implement one weapon. We have to be careful, that all later one needed features are already there in.
    • simple interactive menu: (1) describing the mission goals (2) informing the pilot if mission goals reached (3) informing the pilot he was killed, restart/quit. sound and vision.
      We'll have a look at CEGUI, or have a look at alternatives.
    • very good controls for the spacecraft (this is most important for the quality of the game). Give this task only to very skilled and very motivated programmers that know the game genre. Define one and only one game to be used as a prototype, install it on his/her computer let him play it in front of you and discuss it. This is so important… :D
      It is rather impossible to create a really good steering in just one term. It is going to be some basic implementation which will need much fine tuning in the following terms.
    • simple HUD (showing health and ammo _only_) in a second iteration you may show more..
      We are not entirely sure, if ogre provides what we need, or if we are going to have a look at CEGUI.
    • explosion effect for killed spacecrafts: sound and vision
      It should not entirely be neglected, but there are more important thing to be done first, such as collision, weapons, AI…

Ticket Meeting with Beni and Nico

  • GUI - We can look at CEGUI. The GUI is for the Stations, when the ship docks and for the menu. For this we may want to put two students to work.
  • AI - We have to develop that three layer system, proposed by Martin. We should maybe let two students work on this as well.
  • Weapon System - We need some weapons: energy, rockets and stuff with effects and everything.
  • Ship Upgrade - Our space ship can be upgraded with the weapons, engines and electronic systems. For this we will present a suggestion of what we need and how the parts are dependent of each other. A Scissor-Stone-Paper System should make fights more interesting.
    KISS model, the player should be able to download, install and play the games, without having to choose 2000 out of 5000 different capabilities (quote Nowic). So we need a good entry point for a newcomer, while we want to give a more experience player the possibility to customize his ship, alter-ego, et all to his taste.
  • COLLADA support - We want a better mesh format than the OGRE meshes. COLLADA is the best we can think of. Let's make this work till summer.
  • Missiondesign - I dunno if this should be something a student should work on, but we should produce some missions the player can get and solve.
  • Models - We could use more fighters and stations. We should also follow the question if we should make the stations modular. Meaning a station consists of several parts, which can be exchanged with each other.
  • Physics - We need a easy collision detection and reaction. Should we use ODE for this or implement something on our own?
  • Save and Load - Meaning XML-Im- and Export which is partly done already.
    X3n is already working on it. I will be some sort of saving capability between the missions - or in a later stage, while docked at a station.
  • Since we can equip our ship we have to find that stuff in the universe. Pickups will be necessary.
  • HUD - Not really a lot of things to do, but the radar must work if we want to display exciting space fights.

General Critics

We are creating an open source project, and thus we can use many other open source project as basis for what we want to do. It is rather useless to reinvent the wheel if some else did the same job already. And we need to communicate better - also with other project - to get an understanding of how and what they do, and to coordinate efforts for, i.e ogre ode integration.