Quote:Original post by graveyard filla
maybe im crazy, but i don't plan out a single line of code for my game. and for planning out game features, well i dont exactly do that too much either [smile]. i've been working on an online RPG for the past 6 months, and have been coding and adding things as i thought of them / wanted to implement them. my code is NOT a giant mess, the only globals are instances of my core classes, accessor functions are kept to a minimum... theres only some minor problems im having with data replication in inheritense tree's, but other then that, the core of my engine is pretty slick. im not trying to discourage you from planning out your game / code, but, for me it wasnt required. i simply dont have the attention span to plan out all my code and design [smile].

however, maybe because this isnt my first project i have a feel for the code. ive made a few games in the past (only simple games, like pong and pacman). after programming everyday for a certain amount of time, you start to "see" the code. it's hard to explain, but...

one thing i should note, when i want to add a new feature to the game, i don't just open up the IDE and start coding. i start out in notepad, and right out my ideas in english. then ill write some pseudo code, and then some actual C++ code. only when i have the complete design written down in notepad and everything is connected do i "drop it in" my engine. this has worked for me rather then planning everything out from the begining.

i also know of at least one other person who doesnt plan out their code. they have been very successfull in developing an adventure / RPG game over the past few years, and he plans on making a decent amount of money selling it shareware.

good luck.

well, that might be true... the example above was one of my first "bigger" projects with maybe 4k LOC. better to learn planning no simple stuff before you fail big time. i dont know how it is if you use classes and all that oop stuff. i'm currently catching up on that as well... it seems to make the code much more readable but i guess it cant replace good planning. normally i would say you dont have to plan a project < 2k LOC but its better to get used to it from the beginning.
maybe its just the subject of my studying that makes me think about planning... :|
edit: since you use notepad to write down your ideas you _have_ some kind of planning. maybe not as extensive as i suggested but the only thing it has to do is work. if you achieve good results with the planning you do then its no problem imo. i just wrote all those things to point out that you can eliminate possible error sources from the beginning when redesigning isnt as hard as in the middle of a project. i hate it when projects fail but i always aim too high and i realize this too late. since the time i plan my projects i see those difficulties in advance and can either avoid them or put the idea on the "todo when the time comes" stack.
edit2: and imagine if you work in a team. how do you explain your ideas to your teammates? if you just tell them "you do the collision detection" the results may not look as you thought they would. sure, you can explain everything in detail but by doing so you would also plan. i also know ppl who code without planning but they dont code as a team. as soon as you have to give instructions to someone you have to plan something in advance.
the planning can be minimal but it has to be planned... (just my opinion. ignore me if not planning works for you)