Irrlicht future - compared with OGRE and NeoEngine

[pow]

Hi - I just found you from a flipcode blip - I have been programming for a bit with OGRE and keeping an eye on Neo as it develops - the work here looks fantastic too.

I read the Forums but didnt find much there or on the site about plans/direction for Irrlicht.

First let me say that after using OGRE and reading the API for Neo that the APIs here are much much easier - hurray. Also, I like the integrated collision (can it do object-object collision or just player?). The special effects are great too.

OGRE seems to be focusing solely on being a graphical engine - and the water FX that one developer is working on is simply stunning - if you havent seen it go over and take a look - best i have seen in realtime.

Neo has integrated collision and a first pass at physics code which is looking really good. (any plan for physics here?)

Wondering in general on the direction here - find myself wistfully wishing for the API from here and the FX while having the broader set of modules from NEO and the visual punch of OGRE.

Oh, and to see what I am doing look at http://www.flatland.com - that was my first generation stuff from a few years ago - I am working on a new version that includes multiplayer and a built in block map editor - 3D for the rest of us!

[indianbeard]

Irrlicht seems to be much much friendlier than Ogre.And it seems that
Ogre was in development for a much longer period.The reflection env mapping is good and the rain drops are nice. Atmosphere surely
brings more life visually.

[niko]

Hi,

a bit with OGRE and keeping an eye on Neo as it develops - the work here looks fantastic too.

I also take a look at them from time to time, they are quite nice engines too. Ogre is a very complete engine with lots of cool features, and Neo will get a cool game engine once it is done. You can learn a lot from them. But there are some things why I dont want to use them and considered to code my own engine. I listed the main reasons in here: http://irrlicht.sourceforge.net/faq.html#advantages

First let me say that after using OGRE and reading the API for Neo that the APIs here are much much easier - hurray. Also, I like the integrated collision (can it do object-object collision or just player?). The special effects are great too.

Sure, it can do object-object collision. Simply add a collision animator to the object which should collide.

OGRE seems to be focusing solely on being a graphical engine

This is similar to irrlicht: There is no sound or network code in it. That s because there are dedicated libraries for sound and network, who are able to do this better then irrlicht. And there are lots of libraries, I wanted the programmer to be able to choose which library he will use. But maybe I will create a small reference implementation as addon for irrlicht, with sound and network.

Neo has integrated collision and a first pass at physics code which is looking really good. (any plan for physics here?)

Not yet. Again, I think it would be better to let a specialized physics engine handle this. Like ODE or Tomahawk (or how it is spelled). But as always, there might come an internal physics system in irrlicht - if the users need it.

So about the direction of irrlicht: It simply wants to get a cool library for creating 3d applications (mostly games) with a very simple to use interface. In version 1.0 there will be everything in it for creating 3d graphics in games. It is not really clear what this will be. It depends on what the users of Irrlicht need. For example: Maybe there will be also a level editor.

[ash]

Very nice engine.

If you did a level editor, I would suggest something interactive like the Cube engines editor, you can load a map and make changes while you are in the map!

Keep up the good work, please keep at it until 1.0!!

[Guest]

i would suggest looking at coding your own physics because ODE is a very limited physics engine - its mostly for vehicle sims - i used it for some box stacking and such and it ugly - besides, there are so many game oriented things that physics can do (look at havok) that it would be better to make a subset of them for your engine - unlike sound and networking there are not very many good physics engines out there, especially open source

[Miguel Melo]

i would suggest looking at coding your own physics because ODE is a very limited physics engine

But from my own experience (I coded a physics engine for a car racer many moons ago) physics as a concept is exciting but a b*stard to code. e.g the stacking scenario is a fairly envolved process, what with all the feedbackforces and dependency graphs. Then you also have to add all sorts of bodges so that you can damp the bodies so they don't twitch and take into account that time is sliced into frames so that detection of events is almost always late...

Other than for the research part of it (if you feel that is fun - I do), I would not code a physics engine unless I really, really, really had to...


Miguel

[niko]

Other than for the research part of it (if you feel that is fun - I do), I would not code a physics engine unless I really, really, really had to...

Hey, there is someone in this forum who thinks just like I do. There are other, more important parts missing in the engine, which I'd like to insert first.

[Guest]

there are a lot of things you can do with simple physics (which seems to be the approach taking shape in neoengine) - car physics is a tar pit - and ODE might be best for that problem - but things like stacking blocks, flocking birds and fish, billiards, etc are not THAT hard and add a lot to an engine out of the box - and sadly the only opensource physics out there is ODE

well, anyway, lots to do in other areas of irrlicht but a stab at a physics plugin would be cool - or just an FX plugin structure would be awesome!