Back to UDK...again.

http://www.crydev.net/viewtopic.php?f=40&t=109134&p=1136178#p1136178

Here is a mirror of the post I just made on the CryTek forums. They might take it down, so here's a mirror:

Sorry guys, this pushed me back to UDK (for either the 2nd or 3rd time, can't remember). Thankfully I had only done preliminary work this time. I have not been able to work on my game for four days, so I'm out. I'm washing my hands of this mess for the last time.

You can't consider CryEngine a serious alternative to other game engines if issues like these are still haunting the engine. The DRM is questionable. No other game engine currently controls its users this much. CryTek essentially owns my content....I work on it at their whim. They control access to it. And that really sucks. Case in point I'm going to Eastern Europe this year for a month. I won't have internet. There is absolutely no way for me to work on my CryEngine based project without it. There is no mechanism for me to get a license. There is no way to sign an NDA to say 'I won't steal your software.' All I get is a 'contact CryTek for more details' and details are never given. Yes, I have sent you an email.

Originally I was okay with this. I wanted to work on new tech. But now I can't even access the maps I've spent hours on? It shows utter contempt for your community. Yeah, UDK is technically inferior, but it works. I can make a game on it. 

So I'm left thinking that CryTek is putting this engine out there for some other reason. What I can come up with is a sort of 'me too' approach so they look competitive in the tools market with UDK. It's kind of like a pantomime of having a real game engine, of having a real community, for the sole audience of shareholders and game company execs. 

I'm out. It's a great engine, but unusable for any serious indie game making. This 'maintenance' was the final nail in the coffin. I haven't been able to load one of my own maps for four days running. 

So, if anyone from corporate is listening: stop tying the hands of the people who only want to make games with your software. If you're serious about the community, prove it. Remove your DRM. I'll sign a contract, whatever, but don't stop me from working on my game. Don't hold my art (as crappy as it may be) hostage. Really compete with UDK rather than just using the long hours spent by well intentioned people to cynically inflate your market share.

Logos and Banners

Moddb requires a logo, so I've been kind of making some half heartedly but they all turned out pretty crap until today. The problem was I was over thinking it, and also that most game logos suck. Wolfenstein's own logo is pretty iffy. The newer one isn't bad, but it's got a lot of cliche elements to it, multiple layers with a 3-D outlined logo. It looks pretty generic in the 20xx gaming era.  I was trying to do the same thing, but since I'm not remotely competent as a graphic designer what I was doing sucked. It was just bad.

http://en.wikipedia.org/wiki/DIN_1451 is what I want to base the HUD off of. No Scaleform, no crazy floating widgets, just something clean that doesn't get in your way. Almost vectorized (well, it probably will be vectorized in actuality). I was messing around with the generic layer-fu that you pick up doing game textures and this just popped out:

I know it looks like your generic hipster album from 2010 or so, but I haven't seen any game logo that looks like this. I tried adding the various sundries that modern day gaming logos tend to go for, but everything else made it garish. So, I'm going with it. It's the rug that really ties the room together, if you will. I'm also planning some other ones (classic photos with just the overlay) so I can have a World War 2 desolation flip book of sorts.

Infantry only.

I had someone ask me today why there won't be any vehicles in NovaWolf. The answer is simple: they make games suck if you don't do them right.  The reason is that they're usually vastly overpowered compared to the guys on the ground, especially in WW2.  Ever seen someone take on a tank with just one guy? Sure, it happened with those sock bombs from Saving Private Ryan, but the reality was far different.  The game reality is far different too.  Sure, you have things like UT3 Vehicle CTF and Quake Wars, but that's the future, when you have an array of overpowered fusion based weapons capable of reducing everything to slag. The average Fritz in the Wehrmacht in 1944 only had his Panzerfaust and a clandestine bottle of Schnapps for courage (if he wasn't in some elite anti-tank battalion). Unless I made the Panzerfaust hugely overpowered for vehicles it just wouldn't go. Also, there's what I like to call the "cyclops syndrome" from Quake Wars, where you get that one guy in the ultra heavy armor sitting in the spot that nobody can reach with artillery blasting every single spawn.  I don't want that, and neither do most people. So, say it with me, with pride: "infantry only."

I'm going with UDK, but CryEngine sure is pretty.

For the last year I've piddled about with making a game. That game in question is a port/update/tribute to 2001's Return to Castle Wolfenstein multiplayer.  I've taught myself to model, rig and animate. To do level design and even some coding. However, what took most of my time was the so called "engine evaluation" phase. I've tried nearly every engine under the sun, from idTech 3 to Neoaxis.  All of them (with perhaps the exception of OpenWolf) fall short in some fashion of being able to easily reproduce RtCW in a modern context.

I have this mantra now where I say to myself over and over: "I will not use CryEngine, I will not use CryEngine," because when it comes down to it I *really* want to use CryEngine. It's truly a beautiful thing. Realtime global illumination is amazing.  The things you can do with physics are awesome. Also, the experience feels "tight" and robust in a way that previous engines, such as Unreal Engine 3, just don't. The animations are more fluid, the visual grammar more engaging, and the sense of "realness" more palpable in CryEngine games.  Crytek has clearly done a huge amount of work delivering to the world the first real next-gen engine.  Experiencing Crysis 2's opening scene (the submarine flood) in DirectX 11 mode with ultra textures makes UE3 cutscenes feel like puppet theater.

The problem is, of course, when you actually try to make a game on it. This is not a thing for mere mortals. I have at my disposal no less than three programming books on C++ and Lua for reference. I've had about 6 months of experience programming C and Python and have a basic idea of what is happening when I look at normal game code (i.e. not rendering code).  The actual code in the CryEngine is great. Things are logically organized and, as far as other people have told me, internally consistent.  The problem is there is absolutely no documentation.  The constituent parts are there. You could script an entire game type using just the extant Lua scripts. The problem is that nobody knows what most of them do. Nobody knows on the forum, the documentation is missing the most important files (for me) and the only recent book that discusses CryEngine has exactly 3 chapters devoted to programming.  I could figure it out, sure, if *all* I did was CryEngine for 6 months.  But I have to rig, I have to model, and most importantly, I need to make a game.  I've learned the hard way that nobody else seriously wants to do this except for me, so I soldier on.

So sure, UDK may be clunky and smell like Icy Hot and Werther's Original on account of its advanced age, but every line of code I've looked at so far has had some sort of documentation. Because of the mountain of tutorials available I can visualize what and how I'm going to do this in UDK.  In CryEngine I'd be better off with a Ouji Board and the ghost of Dennis Ritchie.

Please Crytek, hire someone to write documentation and tutorials. Seriously, it took 250 people to make Crysis 2, which was about the most mediocre title gameplay wise I've played in the last 5 years.  Maybe you could devote 2 or 3 to documentation, so someone could use your beautiful engine for something else than heavily scripted monotony.

Beach Version 2.0

One of the things that bugged me about mp_beach was that, well, it didn't really look like a beach. At least not like an Atlantic wall sort of beach. The thing the original map was supposed to convey was the idea of pillbox bunkers and casements along beaches like Normandy.  These apparently had too many curved surfaces for 2001 technology, so we were left with a small bevel on top of largely cubic structures. Thankfully my PC these days can render a million polygons at 200 frames per second, so I've decided to go through the old map and make it look something like the actual Atlantic wall looked some 60 odd years ago. The layout is still roughly the same, but I stalled on remaking the map because it was boring for 2013. Hopefully it should play as good as the original, because there are still servers running it 12 years later.

CryWolf, NovaWolf? SteinenWolf?

So, I'm about ready to seriously start coding and rigging and animating for my Wolfenstein tribute. The assets are done. I have a level roughly completed, along with various props, character models and weapon models ready to go. It took waay longer than I thought it would. But it was rewarding.

The reason why it took so long is that there is *so* much that goes into game making these days. It used to be you could just fire up Radiant and kick out a level in a couple of days, as long as it was basic. Now everything takes 10 times as long on the asset creation side. I have to model it, UV it, sculpt it, bake it, texture it and finally make sure it renderers right in game, which is no small feat given the myriad of options in a modern game engine. Add to this various sundries like making collision boxes for levels. It's a pain.  Yet, some things are way easier than they used to be. Take RtCW and ET coding for example. Everything, down to the number of frames a reload animation takes, is hard coded in C.  I can do half of what the RtCW game code does in Kismet (or Flowgraph) in a fraction of the time that it took really competent coders days to write. You win some, you lose some. 

Anyway, maybe in a month or two I might have a rough alpha ready. Hopefully. It's coming.