10GB upload space fyeah. Gitlab access - Desu!

[Blutreiter]

Entrox was nice enough to provide upload space and wrapped the present in a nice Gitlab. -> git.desu.es

Please PM Entrox for access or post here.

[EntroX]

funfact: only 1 person has asked for access

[Blutreiter]

funfact: only 1 person has asked for access

So gief pl0x. (We're the slowest dev team after the Duke Nukem disaster)

[Pacefalm]

Be careful, you might be getting hired by Valve to work on HL3

[Devec]

I am officially requesting access hereby.

I have been slacking a lot (was away for a week) and I am going to pick up some things again to get art assets going. I will probably make a thread sometime soon about it.

[EntroX]

devec/blutreiter created and sent info via PM.

[NoirAvlaa]

funfact: only 1 person has asked for access

Everyone else is terrible.

Stahpatalk

[EntroX]

out of curiosity, are you guys using this?, if not i'll repurpose that VPS :V

[Corwyna]

Guess the curse of the subforum strikes again. Give it a subforum and it dies...
Shame.

[NoirAvlaa]

Guess the curse of the subforum strikes again. Give it a subforum and it dies...
Shame.

The biggest shame is that this tactic didn't work for the ponies :/

[Blutreiter]

Guess the curse of the subforum strikes again. Give it a subforum and it dies...
Shame.

The biggest shame is that this tactic didn't work for the ponies :/

Heh.

[Devec]

Well it is very unfortunate. Personally I think it was a case of too much brainstorming and too little incetitive, I know I suffered from that. I would still like to continue but then it would be a project where we don't set such ridicilous standards, because the whole idea was really ambitious. If anyone is still interested on working on this I could set up some concepts for a perhaps somewhat more realistic vision. I guess starting with something smaller like a 2d sidescroller like the people behind gunpoint did would make the project perhaps somewhat more feasible. In any case I am still mucking around a bit with 3d and teaching myself more about it just as a hobby rather than something I see being really appropiate for the project.

[helgur]

Going the 3D way was overcomplicating it a wee-bit tbh. I found UDK to be a bit overwhelming and was hoping we at least would get something very simple off the ground I could look at and learn from. Did darkflare even start with the game in UDK?

[Blutreiter]

I never heard anything about progress being made. In retrospect, the UDK decision was pretty fail.

[Nicholai Pestot]

As a professional developer who has been involved in (and is currently involved in) both large scale and indy projects:-


You need 2 to 5 people to churn out a simple game at most. Don't go over that number on your first project or you will fail unless you have a professional project manager/producer willing to kick in some of their spare time.

Don't accept anyone into the team who only has an idea. You can't afford dead weight like that. Small teams are all technical (in some form, be it code/art/music/whatever). Anyone with an idea and little to no technical knowledge is useless to you. Anyone can shit out 20 good game ideas. Idea-only people are worthless. Its the implementation that's hard.

Get yourself some version control software for your files, or accept the project is going to fail.

Tell those 'make it multi-platform or don't bother' people to fuck off. If you are new at this, bug fixing on one platform will be hard enough.

Use unity. No, really. Use it. Nothing beats unity for people who are new to game development.

Be flexible and don't get too attached to your starting work. 90% of the work happens in the final 10% of the project. Almost everything will change.

Nominate a leader and have them set deadlines. People who repeatedly miss deadlines for no good reason (there are plenty of valid reasons, but the guy running the show needs to spot bullshit), need to be booted and replaced.

9/10 indy projects fail, because they approach the project in a haphazard way and burn out after a couple of weeks when the scale of the task becomes obvious. Good luck.

So how much of what I said came true?

FYI the Indy project I was a part of fell apart as well (thankfully I have a day job as a developer to pay the bills). Still, onwards and upwards to the next one. Every failure is a lesson learned

[helgur]

Thats truly some unique insight you are bringing there

In other news the sun goes up every morning, more news at 11

Also, part of the reason the project stalled as I see it, was because some of what you say people actually jumped on and decided to go for, namely unity. Its all fine and dandy if you actually have people who know unity, but when it was decided we should switch to unity no one was remotely competent to use it, and the people who chanted "use unity!!oneeleven" on the forums wasn't willing to pitch in either. Same thing with UDK

[Nicholai Pestot]

That's a pity.

I'm surprised about the development issues you had with Unity. I have worked with rookie developers that have been able to roll with it after a couple of days of watching tutorial videos.

What were the problems you chaps had with it?

Thats truly some unique insight you are bringing there

Please no bitterness at me. I didn't make your project fail, I'm just interested in what caused it and wanted to see how close my own analysis of your situation was. Failure is a learning experience for everyone, not something to be feared or ashamed of.

Successful people are the ones who learn from their fuckups and try 1 time more than they fail

[Blutreiter]

That's a pity.

I'm surprised about the development issues you had with Unity. I have worked with rookie developers that have been able to roll with it after a couple of days of watching tutorial videos.

What were the problems you chaps had with it?

Thats truly some unique insight you are bringing there

Please no bitterness at me. I didn't make your project fail, I'm just interested in what caused it and wanted to see how close my own analysis of your situation was. Failure is a learning experience for everyone, not something to be feared or ashamed of.

Successful people are the ones who learn from their fuckups and try 1 time more than they fail

I believe the reason not to use Unity came from the realization that our game heavily would have made use of lighting and shadows (most important game mechanic according to design) and the free version doesn't handle lighting the way we wanted to, so a Unity license means lots of cash upfront. So we decided to look into other engines.

The only proper progress had been made in Allegro by Helgur (mad props for that still), but voices grew louder that UDK was much better (and also 3D wooo... or something) and in the beginning, UDK really looked decent. Darkflare said that he had experience in UDK and whipped up a proof of concept just like that. We were sold.

So now it came to actually implementing the random map generation we needed as a base. Reference material was gathered, discussions were had, several Skype meetings went by and we wanted DF to get us some initial code base we can work with. "No problems" and "Yeah this weekend we'll do something" all around.

Never happened.

[helgur]

Please no bitterness

It's not biterness, I am just expressing annoyance at your presumptiveness and "told you so" arrogance and also you pointing out things that are obvious. Also what blutreiter said about unity.

[Nicholai Pestot]

Ahhh, the dynamic lighting.

I would have rolled with that in Unity and used a manually created range culling server-side to limit entity creation for players. If the project was succeeding, I would have then gone for a fundraiser to try to get the pro version and overlap some dynamic lighting on top of the culling system.

With your game environment it would have been (relatively) easy to dynamically break it down into zones, with entities subscribing to each (room/corridor) zone as they leave/enter.

Subscription changes could be checked for every 1 or 2 seconds (rather than every frame) to limit processing overhead.

After this you perform a direct range check (limited by zone subscription adjacency – you only check range culling for a player against entities in their zone and adjacent zones).

After that you further break it down with LOS checks on anything that passes the distance check.

Effectively three tiers of culling, with each successive step working on a smaller set of items so that it can be more accurate without being too computationally expensive when entity numbers increase.

Only the stuff that passes these three levels of culling on the server gets sent to the clients.

Then you overlay dynamic lighting on the client to make this culling look good. It's one of the last things you worry about.

Yes. I know. 20:20 hindsight and all that. It's easy to come in after the fact, but difficult to apply while doing, especially if doing while learning.