It looks fine, but this tells us nothing about the engine. It could be designed horribly and we would never know. Nevertheless it's clearly functional and I'm interested to know how long this took to develop.
For large projects I never request source code, I know a lot of people have a problem with disclosing it (although I'll never understand why). I just wanted to know roughly how long it took.
I've been working with SDL myself lately and I find it quite interesting, especially the sound side of things, you really have to work to get a sound component with SDL.
No worries, I don't have a problem disclosing it, I just don't want to release unfinished code. If someone wants to use it and it doesn't have all my fixes, I would think it would look poorly on me.
(edit) for instance, I have this *really weird bug, if I make more than 5 blocks in a row and make one of the blocks in the middle move up and down, the downwards collision detection won't repair the player's velocity, so he falls through the blocks... [pulls out hair]
The creation of *this* engine began on March 9th.
I love working with SDL, and I really haven't worked with sound at all!
There's a seperate library called sdl mixer that you can use for sound but if you feel like playing with some low level programming give sound mixing a go on your own. Being able to mess with raw data like that is the main reason I love C++
@blackcoder Thanks! The song is called "Chrono Trigger Corrupter of Time OC ReMix"
@quirky I've seen the mixer lib (on lazyfoo) and I love music, so both definitely going to be looked into! Thanks, by the way, I really appreciate your input!
Oh, Have you had any experience with pathfinding in a sidescroller? Anyone?
As it has been said, this video doesn't say much about the engine itself.
It looks nice, but the thing I like best is the idea to edit map at runtime. I remember when I did things like that, I would write the core game, create a new project for map editor, write a (simplified if possible) map class, some basic GUI, read and write functions for a map file, create the map and finally load it from the main project. And of course each time I added some new feature, I had to update all of that. That was a bit silly of me..
What kind of path finding does a sidescroller need?
Pathfinding in side scrollers? He said he wanted to make an rpg, so I'd assume it's for the enemy AI. Would be pretty stupid if they kept jumping into holes.
Currently I'm thinking about making a small file interpreting pathfinding algorithm. I could set everything in a map editor, but I don't know... I would much rather have some predetirmined AI personality base which would be able to make it's way around the map without nodes and such....
Well, for pathfinding there is A*. You'll have to use a smarter function to determine the neighbors of one tile than you usually would (to deal with gaps, pits, walls and jumping)
I have read quite a bit about A* it's probably what I'll end up using, unless I buckle down and preform stupid acts of code. I maaay just make some random movement with event systems for each AI. Perhaps something like move in a direction until one of the following events happen: See player, hit wall, stand by hole, etc...
Seems doable. However I think that an existing library would be much easier. Gah!
I have never seen a pathfinding algorithm really different from A*. You have a net of nodes, some available, some not and it's A* already.
The pathfinding part itself is trivial. The hard part, as I mentioned, will be finding neighbors of a node and you won't find a library to do that for you.
Indeed, I have absolutely no problem implementing it myself =)
Aaaanyways, besides that I'm going to have to upgrade my camera class. Currently it has the ability to follow objects, and smoothly move towards them. However it's still a little jumpy if I allow my player to teleport. So I'll have to make some sort of panning teleportation thing or something. There's an awful lot to think about!
So, basically your camera locks on an object, and if it jumps the camera jumps right afterwards? You could set something like a max speed for the camera, and rather than letting the camera jump let it actually follow the object with it's own speed (with a set max speed, or a max speed depending on the distance to the locked object).