Not much to say.  I thought it was nice, simple entertainment.  (Edit for clarity: I had a lot of fun - previous sentence intended as compliment; "simple" = good thing).

Don't suppose we could get ladders or something?  Maybe it's part of the metaphoric thrust that you're perpetually going down, but something in me would appreciate being able to explore in any direction. 

I really don't see the problem about making your own stuff, most of the times it's not reinventing the wheel, because there are not complete tools of development out there if you're not speaking about the Source engine or U3.

Sorry, I've lost track of what I'm arguing here. 

Is it the question of whether a general purpose 3d engine is possible or desirable for a lot of projects?  Because I think the existence, use and success of commercial engines provides good evidence that that is the case, and that's why I brought them up.

Or are we arguing about whether there exists, now, an open source engine worthwhile for using in the kind of game it seems like is being built here?  On this, I happily concede that maybe you're right.  Maybe there isn't anything worthwhile to use.  I guess my point would be that there appears to be a tremendous number of people who've made an attempt, and at very least it seems like it would be a good idea to see what they've done before writing your own general purpose 3d engine.

Or are we arguing about how much value a good engine/toolset provides in development, and how long it takes to produce this kind of code for a new project?  In this case, I'd say it depends on the project.  You say 1 or 2 months (and I assume you're meaning 1 person part time or something).  I think that's possible depending on a game, but I think an engine developed in that time frame is going to put heavy limits in terms of how well you can produce and use art and other content, as well as limits on functionality that might be desired down the road.  I think this is especially the case when what you're doing essentially mirrors average commercial games: outdoor setting, animated humanoid characters, etc..

There's obviously a million things that vary, and the right path really depends on what you want to produce and what you enjoy doing.  I'm just giving advice based on my opinions and values. 

However, maybe this puts it in perspective.  From the original post: "We are a team developing an 3D Engine for 2 years and a half". 

After 2 and a half years, there's not a demoable game.  This is how 90% of the independent 3d projects I see go (obviously that statistic is off the top of my head, but I think it's about right).

Also, as a last point, the last time I did a 3d project (long ago!) I wrote my own engine - so my position isn't some bizarre absolute. (See www.jumpmanzero.com).  In defense of my mental stability, I knew exactly what I'd need for the game (which was nothing - no animation, basically fixed camera, no shaders, procedurally generated models so no import) and a large part of my interest in developing the game was learning about DirectX.  I ended up spending a few days on engine (not 2.5 years) and the rest of the dev time (a few months of scattered hours) doing game content (the part where an independent developer has something to bring to the table).  It ended up being a successful game for me (lots of downloads, lots of positive feedback from players).

If someone wants to write their own engine code, then that's a perfectly reasonable choice and they're entirely within their rights to do so; they really don't need to justify themselves to back-seat programmers on the Internet.

Sorry, my understanding of the "feedback" forum is that people wanted "feedback" on their game, how it's being developed, and what not.  That's what I gave.  He's obviously under no obligation to listen to anyone's suggestion or justify himself.  If someone asks for advice on how to get somewhere, and I say "I think you should turn left" does that make me a horrible backseat driver?

And yes, if someone wants to write their own engine, then that's a perfectly reasonable choice.  But if what someone actually wants it to make a good game, my contention is that writing your own engine may or may not be a reasonable choice.  And the last part - "within their rights" is just silly and offensive.  Obviously I'm not contending that it's not OK for someone to write whatever code they want.

That's not true at all, there are very few engines that are really usable, maybe Ogre and something else...

That seems fairly dismissive of a lot of code and a lot of people's work.  If we include commercial engines (like Unreal, or more hybrid type things like Torque), then it seems even more of a reaching claim - and overlooks the fact that there are many commercial or otherwise successful games made with other people's engines. 

Really - why would I believe that my engine code is going to be any less bloated or unusable than the work someone else has done (and taken the bother of posting and/or specifically refining to be useful to others)?  Again, maybe I'm the one to finally get it right - but everyone can't be right here.

Now I don't expect you to try out hundreds of engines... but I guess that's part of the problem.  I don't expect anyone to try out all these engines or collaborate and share code on this front much at all.  I expect that a large percentage of independent 3d game developers will go ahead and spend most of their time fiddling with engine and tool code, and, in the end, produce either nothing or a game that ends up being a boring, low-content tech demo for their engine.  And I expect the primary reason for this is, as before, a typical programmer's (including me) built-in preference for coding over reading documentation. 

Imagine an alternative: people actually get together to improve an engine and build a selection of common tools around it.  Every developer that uses that start ends up giving back something.  In the end, you can produce an interesting game using conventional 3d graphics very easily and still have the flexibility to implement odd features as required.  That won't happen if everyone starts from zero.

Professional game developers DO their own engines for the needs of the team and the game

Sometimes they do, sometimes they don't.  Sometimes they do make their own engine, but probably shouldn't have.  I've played plenty of games that could have been implemented in established engines, but instead were crappy bugfests in a house-built engine. 

Commercial games also have huge budgets, huge teams, and often are pushing technical frontiers in terms of graphics.  Most independent 3d projects are exactly the opposite - they have no budget, tiny teams, and are generally not technically demanding (to be technically demanding in a 3d game usually means lots of art resources, which aren't there because there's no budget). 

That's not to say there aren't plenty of exceptions. If you're really doing something outside the ordinary that will require something a "normal" engine won't do easily, then fine.  But if you're developing a fantasy game involving characters with canned animations running around a simple outdoor landscape with buildings and trees, then you're firmly in the realm of things a pre-built engine is going to help you realize.  And by using that pre-built engine, there's a good chance that "next feature" you want is going to be a lot easier to get than it would have been otherwise.

Going to an extreme, consider Defense of the Ancients.  It's a hugely successful game, and it's that way largely because the developers involved had nothing else to focus on but gameplay.  Everything else: engine, models, animation, network code, sounds, is re-used stuff. 

Sure in some sense they could have produced something better by starting from zero.  But could they really have?  How many independent games have started at zero and been as successful or high quality as DOTA?  Not many, because for most games the time that could have gone into making a good game instead went into producing the things the DOTA devs got for free. 

I'm certainly not saying everyone should make mods.  What I'm saying is that it can be worthwhile to concentrate on the game you want to make and take the shortest path. 

Responding to zweifel: But what if instead of putting some scripts, we just develop a system of passing messages between npcs.

My comments were regarding building your own 3d engine; building the AI for your game from scratch is a much more reasonable prospect, and I think having an interesting AI is a great place to really set your game apart.

It's hard to tell with the amount of natural aliasing - but I assume based on the layout that you've gone with the "period appropriate" raycasting rather than now-conventional polygons.  Similar raycasted games worked on the GBA, but I assume this wouldn't work processor wise on an actual Gameboy?  Or did you actually try setting some limits in terms of computation?  

Edit: Whoops, I fail - didn't read the readme.

    - Fantasy style
    - Platform: PC
    - Entirely destructible environment. (destructible trees, buildings, people and so on)
    - Modifiable terrain
    - Entirely modifiable game history (every game player will be different, because of how the game is designed)
    - Lively AI

I enjoy writing "engine" type code, and in work I quite often reinvent wheels because I enjoy doing it, and because I prefer time spent writing code to time spent reading documentation.

But I still don't understand the apparently universal compulsion to write a 3d game engine.  I see 297 engines here: http://www.devmaster.net/engines/list.php?order=date_added&ordertype=DESC (though some don't appear to be engines really) with every possible permutation and combination of features. 

I'm sure you'll learn a lot in the process - but the ideas you seem to be interested are all in "game land", so why don't you spend your time there?  Maybe you've decided you can't implement your game in a current engine.  I don't see anything in your description to suggest this is true. 

Contemplate: likely a good percentage of those 297 engines were created by a developer much like you, who imagined they'd be the one to finally get an open 3d engine right with just the right feature set, and the engine would be re-used by lots of developers and would be the start of a ton of great new interesting games.  While we can imagine a few scenarios that might make several of them right, we're still left with the reality that most of them were wrong and would have probably produced something much more interesting if they'd started with someone else's engine.

If you trade the time spent developing another engine for more time writing an interesting game, I think you'll come out on top.  People like playing games with interesting new ideas, but those interesting ideas are hard enough to implement without re-inventing everything. 

I quite like the game.

I think in some cases the level design is a bit loose.  I think this works later on in a game, but maybe towards the beginning the levels need a bit more structure - a bit less open to get people used to the mechanics.

I don't mind simple graphics.  I think the sprite for your character works good.  The backgrounds and blocks, though, need a bit more character.  Perhaps if they looked kind of "sketched" or something (or MSPainted).  Or just draw them with bolder lines and less internal structure.  Or get someone to Photoshop filter them until they look a bit less like they do now.  Anything.

Sorry - but back to the beginning, I really thought this game worked well, and I quite enjoyed playing it.

I gave it a quick go - really not a fan of the controls.  This is just personal preference - but I hate what feels like a lot of acceleration or animation or something between input and character responding (I'm a big fan of the Gameboy Castlevania games - all of which have very tight controls). 

I hate the flat "one size only" jump arc.  I guess it controls like Prince of Persia (the 2d game).  I don't like that, but other people may.

There seems to be key jamming problems with the default key set (I'm not sure how concerned you are with that, or to what extent you mean to target keyboard gamers).  I had trouble in particular with crouching and attacking cutting each other off (I don't know how consistent this is between keyboards). 

In several cases, tutorial hints popped up when I had already started running away from whatever triggered them.  So the hint disappeared before I could read it, and it didn't come back.  I don't know if I missed something important.

For game balance, maybe it makes sense to start off with an attack that has such a small range - but it's frustrating, and combined with the sluggish control will turn a lot of people off of the game before they have a chance to see much of it.

All that said, I like the genre, I like the graphics, I like the general approach, and I think it's got good potential.

I'm jmzero.  I work as a programmer at an insurance company.  I've been doing hobby game development since the C64 - no big, serious projects but a lot of working little games.  I'm particularly interested in algorithms (I spend a fair bit of time at topcoder.com) and value good game design and balance.

All time favorite games: Starcraft, Star Control II, Ultima V, Jumpman.
It appears this thread grows at an alarming rate, so I'll cut myself off there.