pouët.net

Getting past the 'coding random crap' stage?

category: general [glöplog]
I'm not sure what I exactly want to ask, or what I even want to accomplish with this thread, but I'm a bit bored so here it goes anyway...

I'm wondering how you move from being able to display several decent looking but totally unrelated effects or scenes on your screen, to a coherent demo? When it comes to this sort of thing it's almost as if I have no inspiration at all. Since I cannot model or draw I have to rely on procedurally generating interesting geometry. Admittedly, I'm still in this phase where I try all the cliche things like torus knots, metaballs and tunnels :D... these might be not too useful on their own anyway so maybe that's part of the problem too. But I know that even if I'm able to create a single scene that's very nice and without all the cliched stuff, I'd have no idea how to take it from there.

Of course I could easily entertain myself a little longer with implementing shadow maps and more post-processing effects, but if this code ends up rotting on my harddisk I don't really see the point. I'd like to create one demo that's fairly decent at least before improving the tech so I make a better looking one.

Has anyone else struggled with this?
added on the 2009-05-27 12:52:33 by spinor spinor
only for about 13 years...
added on the 2009-05-27 12:53:37 by Navis Navis
Navis, I just felt like noting: When you started your struggle I was 5 - 6 years old.
And now, this made me feel like shit :).
added on the 2009-05-27 12:56:12 by decipher decipher
Navis: despite your struggling you seem to be getting things done pretty well! hehe
added on the 2009-05-27 13:03:23 by spinor spinor
spinor: that's where talent (as in, demomaking-related talent) is, and that's what usually makes a difference in demos.

At least, today, where the actual coding (except for intros) doesn't really impress anymore
added on the 2009-05-27 13:09:07 by Jcl Jcl
This damn talent! I really want to believe there is also at least some skill involved that can be cultivated if you put in the effort.
added on the 2009-05-27 13:14:13 by spinor spinor
Join a group with a graphician-designer, he will have good ideas that you don't even know how to code :)
PulkoMandy: yeah, if possible I suppose that's the best route to take. Actually, one of my friends is a very good designer, but I cannot convince him to do demos as he feels he cannot model (he only does 2D work) even though he is pretty good at it. But other than him, I don't know any graphicians and/or designers...
added on the 2009-05-27 13:19:03 by spinor spinor
Demo can be good even without any 3d models.
added on the 2009-05-27 13:22:41 by pommak pommak
code a tool, so that you don't have to recompile when you need to change colors/textures. that makes it much faster to get somewhat decent graphics.
added on the 2009-05-27 13:22:41 by neoneye neoneye
Maybe patience and insistence. And then you've finished your first demo which is a collection of effects (it doesn't need to be the best at first). Then you repeat the process for 10-20 times till you starting getting bored and prefer coding the effects than connecting them in a so called demo. LOL. Or you suddenly start to like it. I don't know. It's hard to be motivated.

Maybe a team (if you ever find a good one with the same enthousiasm as yours, if your activities can be synchronized). I don't know if building a good demo framework with some functions helpful for scripting parts and transitions and loading images/objects and other trivial stuff. I haven't created a scripting engine yet, always did the flow of the parts/transitions by hand, always the ugliest code of my demos, and after every release I promised to myself to make a scripting engine or some functions to make this process easy or elegant. Some other people told me it's not necessary in demos. I don't know..
added on the 2009-05-27 13:23:04 by Optimus Optimus
Then, by all means, do stuff in 2D! Almost no one does that anymore, so you have a chance to be unique.
added on the 2009-05-27 13:27:28 by trc_wm trc_wm
Quote:
Almost no one does that anymore


i wouldn't go that far
I guess I could try to convince him to do 2D work for a demo, but it would still be difficult due to a possible bad taste in his mouth from our previous failed game collaborations. (The first one failed because I kind of overestimated my own ability when I thought I could manufacture our own custom motion capture suit :D, and the second one failed because he decided 'no more 3D for me' even though everything went really well.)

@ Optimus: I've already done this in the past (demos which are collections of effects), but years later... everytime I watch them I feel like curling up in bed, all crying and trying to feel comfortable again by masturbating myself to sleep. ;)
added on the 2009-05-27 13:48:08 by spinor spinor
passion and vision, patience and hardwork.. nothing comes naturally most of the time. Demos have no set formula, if you are artistic you follow no predefined rules, never ever conform to any trends. When in early stages always bite but never ever copy. You will find the way..
added on the 2009-05-27 14:11:41 by dv$ dv$
:confused: in the "all-time top" demo list, only starstruck/TBL has a non-trivial amount of 2D graphics/artwork. The rest is 3D modelling or raytracing...
added on the 2009-05-27 17:43:14 by trc_wm trc_wm
trc_wm: You mean non-trivial amounts of cliparts ripped from external sources?
added on the 2009-05-27 17:49:03 by kusma kusma
Who cares if a demo is coherent? To me, the more random, the better. Demos are about eye candy, coderporn, boobs and fun. Concept demos suck because once you grasp the concept, it's just more of the same with no surprises.
added on the 2009-05-27 17:54:27 by cruzer cruzer
I'd say there's 2 ways to approach a demo:

1. Making tons of stuff, then wondering how to make a demo out of it
2. Planning a demo, then figuring out what stuff you need to make for it

2 will result in a pretty coherent, quality demo. Only catch is that instead of lots of fun experimental coding sessions, you now have a big project, tons of boring stuff to do, and you'll still be working on it long after you get bored of the whole idea and wish you'd done something else. I think being able to beat that is as important as talent :D
added on the 2009-05-27 17:54:43 by psonice psonice
psonice is wise. I'm currently stuck in 1. but maybe there is a 3.

3. Pick a party, hit the deadline. No matter what. If its incoherent, invent some names and claim each section was written by a different leet coder.
added on the 2009-05-27 18:19:53 by auld auld
its time for the revival of the MEGA demo :-)
added on the 2009-05-27 18:44:58 by magic magic
please wait while loading part 5 :-)
added on the 2009-05-27 18:45:17 by magic magic
As auld put it:
3. Pick a party, hit the deadline.
added on the 2009-05-27 18:47:58 by rale rale
Quote:
2 will result in a pretty coherent, quality demo. Only catch is that instead of lots of fun experimental coding sessions, you now have a big project, tons of boring stuff to do, and you'll still be working on it long after you get bored of the whole idea and wish you'd done something else

...and/or, that several months in, you realise, that the idea was not that awesome anyway or that what you set out to do was not really feasible. ;)

My problem is more a rather short attention span: After two or three weeks, I _will_ get bored and lose focus.
added on the 2009-05-27 18:50:10 by tomaes tomaes
Yeah.. number 3. Get something finished, even if it's rubbish, then you can learn from the experience move on to the next thing. Getting feedback from the community will help your motivation a lot. And once you've proved that you are capable of finishing things you should find it easier to convince other people to help you out with the areas you are not so good in

Also, you don't have to come up with an amazing story before you can make a coherent demo - try something as simple as picking a nice colourscheme (kuler.adobe.com) and using that in all of your cliched effects
added on the 2009-05-27 19:01:28 by evilpaul evilpaul

login