Progressing in small steps is definately the way to stay motivated and focused. If you plan to immediately make a game on the scale of todays commercial titles you will quickly become disillusioned.
Instead formulate smaller tasks which can be put together to make something bigger. Continuously build on previous work. Just seeing something new added or improved will continue to give the team motivation.
erich’s comment regarding design and task breakdown is key to developing any project. Break it down to the bite-size pieces.
But are their tools to help us manage that online? What this type of endeavor needs is one of those project management systems with an admin and a couple of managers to manage the tasks (like NetOffice).
But a variant of this tool might include the ability for anyone to self register and look at the task list. “Model a crate” - say, i can do that! So the new user signs up to model the crate. He gets the specifications for the task and when complete, uploads his blend and a render to that task as an asset.
The manager can then review the assets submitted for the “Model a Crate” task. He can pick the best one. Perhaps the community votes on the submissions (if there’s more than one).
All this would need to be wrangled by a PHP slinger (sadly i am not that.) To customize one of the open source project management systems (PMS) might not be a difficult task.
BCGP can then incorporate this tool and benefit from a tighter team and community. I do like how they present screenshots on the homepage…
But a variant of this tool might include the ability for anyone to self register and look at the task list. “Model a crate” - say, i can do that! So the new user signs up to model the crate. He gets the specifications for the task and when complete, uploads his blend and a render to that task as an asset.
Yes, that’s exactly the type of system bcgp will aim to create. Ideally that would be the best system to present open assignments, completed ones, and a fully organized structure where all progress for the game can be seen. A custom module will have to be created to implement all this later on. Better structure organization will be top priority after the 1st game is done and hopefully one of us will be able to program a custom module like that later on.
Thorgal - Wow, that is extremely specific to fill out. It would be a great help in future organization of teams, but not many people would be able to fill that out.
Rule of thumb to see what someone can do is…
Reputation in Community OR
If they can point to previous projects they’ve worked on (Team or Individual) OR
Let the person dive right in and see what they can do
Cause you never know how fast a person might be able to learn.
The problems of blender game teams are common to all independent game teams, mod teams, professional and amateur collaboration and businesses. Here are some solutions that have worked for others:
Co-operation - Most independent game projects depend heavily on the charisma and recruitment efforts of the founder(s). A cooperative organization (worker co-op?) could increase the autonomy, initiative and sense of ownership of all the members. http://www.ncba.coop/abcoop.cfmhttp://www.wisc.edu/uwcc/info/i_pages/prin.html
Free Software (“Open Source”) - Obviously Blender is already Free Software, but why not make an entirely open source project? While a proprietary/copyrighted project may attract more young developers who hope to “get in” to professional game development, a non-proprietary/copyleft project could attract both young and experienced developers who are interested as players and artists as well.
Free (Open Source) Middleware - Make all of the difficult behind-the-scenes part of your game (especially python scripts) open source while keeping the face of the game that players see copyrighted for maximum commercial exploitation (art, story, music…) This way you can work together with “competitors” that are also producing free and commercial blender games. A win-win scenario instead of working against each other (a zero-sum game.)
I started making Quake mods in 1996. In 1998 I downloaded Crystal Space (LGPL 3D engine) hoping to find a quick way to make 3D games. Instead I ended up eventually becoming a CS developer and actually wrote the first implementation of a few features that were missing at the time - stuff I never thought I could do when I started.
The only thing Blender needs to compete with the latest game engines is work. If you can write python scripts, you can eventually learn how to make improvements to the game engine itself, though you might spend a few months just learning how to compile the darn thing
Very true, I am learning python now myself for that purpose (got my brand spanking new edition of “Learning Python” from O’Reilly!) We should get some more ideas for another community game, and thjen implement these steps and ideas, would make a really great game that would be easy for lots of people to contribute.
Since my main computer isn’t usebale anymore dual the motherboard instability, I thought why not spend some time to learn c++ since blender runs slow on my moms laptop (it has an awefull crappy videocard), maybe some more people here are willing to learn c++ too.
We could fuse as a group and start learning c++ together, get everyones idea, find together solutions and so on. I know that this really helps, I did this iwth Nor.J with learning python and socket. Maybe c++ turns out really wel this way.
I still need help learning python! I understand the language, as in the variables and loops and such, but how does one turn that stuff into scripts and programs? I have had the same problem with every language I have tried (c, c++, java, basic, a few remote others) I am trying to stick with python, instead of giving up like the other languages, because it seems simpler, but it is very confusing. I am very good with computers, but I am not “computer minded” so to speak.
Note: So are meaning learning c++ to improve the game engine or just for kicks?
I am a student game design&development (from HKU, high school of arts Utrecht). I must admit I havent worked with blender that much, though it was the first game engine I got to work with.
I do know of all the problems people can encounter during such a proces however.
A nice idea for genre could be something like an adventuring game- well this idea came from my own passion- but adventuring games dont need high speed, and dont need the bestest of bestest graphics to compete (look at monkey island 4, for example, its quite lowpoly, but does that matter?)
you do need a good level designer and a very good dialogue writer though.
I just put my eye on Crystal Space, but the last time that I did something on c was 16 years ago and it was on c not c++ so I basicly will learn from 0. But I think it is worth the trouble. I realy like the GE of blender but after doing some test with ge and also python, I do not see how can I achive a good game on it. Is to slowwwwww for processing. Please tell me of one good profesional and comercial game make on the GE of Blender ??? Because if is posible for someone to do it, the rest of us will have a hope to do the same.
I agree with that Lemmy, there are aso much possibilities. I’ve got like 500 blender game engine sources here, that each contains an another test or technic to accomplish something. From simple script only mouse looks thingies, to minimaps, inventory scripts, flying and normal fake ai (non python and python) even some mesh import scripts for realtime.
And still people of the community are finding out new technics for games.
HEY , DONT GET ME WRONG. I realy like the ge of blender but I saw my self doing and expending a lot of time trying to do scripts to do the things that other ge can do. Like trying to reinvent the weel. When the blender ge will be as fast as CS or Ogre??? or can manage more than 5 armatures with out having some slow performances? If I do a script to do shadows or dynamic ligths, is this going to be as fast as the others ge’s? I dont think that python is going to to reach c++ any time soon. So thats why I’m looking for some alternatives
The blender GE has 3 major problems, and RCRuiz mentioned all of them:
Re-inventing the wheel. The blender GE needs more features - either built in or through free scripts and libraries. Blender users like myself can help by sharing their creations and discoveries about how Blender GE works. Hopefully this will lead to more direct involvement in developing the GE itself.
Performance CAN be more efficient (especially Armatures.) Some quick experiments seem to indicate that the amount of time spent in “Logic” due to armatures is proportional to the number of key frames in the current action. (not just the key frames immediately before and after the current frame.) This is either a programming error, or more likely a very innefficient design.
Python will never be as fast as C++. The advantage of Python is that the engine works harder so the programmer doesn’t have to. Python scripts can be written and repaired faster, meaning the game does more and gets finished faster. Pygame.org has more info about commercial games that use python. However, this also means that the GE has to have built-in features for all the features you need where Python is too slow.
RCRuiz: Really? Let’s see:
I got about 18 complex(16 bones each, w/ 15-keyframe animations) armatures, fully rigged with 1000 poly models each, on-screen at the same time, and I got around 20 fps. Plus, I only have a geforce2-based card.
RCRuiz: Really? Let’s see:
I got about 18 complex(16 bones each, w/ 15-keyframe animations) armatures, fully rigged with 1000 poly models each, on-screen at the same time, and I got around 20 fps. Plus, I only have a geforce2-based card.
mmmmm, let see how many fps I could get working in c++…mmm… or bether, if I put my limit on 20 fps how many armatures more than 18 can I put on my sceen to make it more atractive ore challange . . . . . mmm . . . . I bet A LOT.
I do not say that the effects can not be done but when those efects are excecuted are not going to be as efficient as a pure c++ game . dont you think
The armature code IS in C++. (When you enable profiling in Blender, it shows python, armature animation and logic bricks combined into a single Logic percentage, so you have to experiment to figure out which is slowing you down. Logic bricks are quite slow, python is faster and armatures shouldn’t be as slow as they are
16 bones doesn’t leave much for face or hand movement. 20 fps is not very good performance. Armature performance is probably determined by CPU, not graphics card. (though having a slow graphics card could create a bottleneck.)
I should do some more testing and post a bug report or something, but I think the armature performance could be better because it should be proportional to the number of bones (or bones that move during the action) not the number of key frames in each action.
(Our armature has 80 bones, about 25 of those bones move in a typical animation. I have a 1 GHz Celeron CPU and GeForce 2 graphics card. Each character seems to slow the game down by about 4 milliseconds - 0.004 seconds of “swap” time - Cutting out polygons has little effect on performance, but cutting out keyframes increases performance proportionally. So the 50+ bones that only have 1 keyframe per action have much less impact on the framerate than the 25 bones with several keyframes per action.)
