ALL BLOG POSTS AND COMMENTS COPYRIGHT (C) 2003-2020 VOX DAY. ALL RIGHTS RESERVED. REPRODUCTION WITHOUT WRITTEN PERMISSION IS EXPRESSLY PROHIBITED.

Friday, October 16, 2020

Four game designers, two interviews

My recruitment efforts for high-quality additions to the DevGame site have already been rewarded beyond my reasonable expectations, as two veteran game developers have already agreed to become regular contributors. Restitutor Orbis, a game designer who is the newest contributor to DevGame, interviewed Chris Crawford, legendary game designer and founder of the Game Developers Conference, in 2005:

Eastern Front (1941) was one of Crawford’s most noteworthy creations so I decided to press him for details. “Eastern Front was a creative implementation of an obvious idea. ‘Let’s do a good wargame on a computer!’” he said. “Pulling it off involved an awful lot of creativity, but it required tactical creativity as opposed to strategic creativity.”

I was puzzled by what he meant. Crawford has a reputation for being outspoken, but it’s a cryptic sort of outspokenness, profound to the point of incomprehensibility. Talking to him can be like reading A Brief History of Time at 120 words a minute. You always feel like you’re missing something.

“Tactical creativity is implementation creativity. How do we build a good map? How do we move units around? How do we build a good AI system? You already know where you are going and you are just figuring out how to get there.”

“So would you say in today’s game industry we have a lot of tactical creativity and less strategic creativity?” I asked.

“Nowadays the stuff we call creative is tiny, tiny stuff. It’s hard to even call it creative at all. Technically, yes, I see a lot of creativity. But I see almost no design creativity in the stuff that’s coming out there.”

What was beginning to become apparent in 2005 is now completely obvious to everyone 15 years later. Read the whole thing. For my part, I interviewed Brad Wardell,  designer of Galactic Civilizations and publisher of Sins of a Solar Empire and Sorceror King, as part of the 2016 DevGame course:

VD: You've moved from doing science fiction with Galactic Civilizations into doing 4X fantasy with Fallen Enchantress and Sorcerer King. What were some of the challenges that were involved in moving from science fiction to fantasy?

Brad: The biggest one for us was going from a space-based game like Galactic Civilizations II to a land-based game like Fallen Enchantress. Specifically, the terrain. You are dealing with the ground. And that turned out to be a huge challenge for us because we had never had to deal with it before. We had never really run up against things like video memory or the limitations of DirectX in terms of how to make a mountain. You think about it, of course, but how you make something like a mountain can be limiting based on DirectX, because there's only so many points you can put on there. So that turned out to be a huge hurdle for us, and that really bit us in the butt, because, at the time, we didn't do our homework on what we could and couldn't do with the current technology.

VD: Interesting. That's very timely because we're going to be getting into things like polygon count and so forth when we talk about art later today. Now, in the publishing world, the market for fantasy is considerably larger than the one for science fiction. Is that true in games as well, or do you find that science fiction usually outsells fantasy?

Brad: I read mostly science fiction myself. In the game arena, I would say science fiction tends to be a bit ahead of fantasy, only because the problem people run into with fantasy is that they think fantasy means medieval Europe with magic. And that's not a just a problem in terms of the designer's limits, it's more the expectations of the public. If you move too far outside the box, you are punished for it in the marketplace. Whereas in science fiction, you have a little bit more room to breathe.

Even if you're not a game developer, or a wannabe game developer, there is a good chance you're going to learn a lot of interesting information from these interviews. And if you're a gamer, you're definitely going to want to add DevGame to your daily bookmarks list. 

Labels: , ,

31 Comments:

Blogger Azimus October 16, 2020 5:15 PM  

The SDL has been prolific today

Blogger RedJack October 16, 2020 5:25 PM  

GalCiv,SOASE, and SK are all favorites of mine. I have been playing a GalCiv game since.. well I don't want to think about it.

Sins was a great game too. Many a transition to night shift spent playing it.

Blogger TroubleSpeak October 16, 2020 6:10 PM  

I really hope Wardell is working on another SOASE game. Such an excellent application of 4X in real-time gameplay.

Blogger Yossarian October 16, 2020 6:45 PM  

This is great. An old friend recently got in touch and asked me if I wanted to make a game. Then Nvidia releases its 3000 series. Then AMD announces their Ryzen 5000 series. Now this blog pops back up.

Things are moving so fast in this realm that soon enough hardware limitations will be a thing of the past. Devs will no longer have to worry about optimization and they will only have to focus on creation.

Blogger Unknown October 16, 2020 6:50 PM  

I know nothing other than being a player and referee.

I am interested how Traveller went from being the greatest scifi game with a universe slightly less fleshed out than the Forgotten Realms, to crashing totally off the rails.

Blogger VD October 16, 2020 7:02 PM  

What do you mean by "crashing totally off the rails"? I have some ideas about what you might be talking about, but I'm curious to hear your perspective.

Blogger VFM Bear October 16, 2020 7:48 PM  

I've always liked Brad and Stardock. When I was having problems getting GalCiv III to run after it first released, one of the dev team answered my bug report and asked if he could call me.

So I got to spend an hour chatting away with one of the dev team about games we liked (we were about the same age, which is to say about the same age as the SDL) as he remote accessed my computer to look into the reason behind the failure of the game to start. My Steam library was open and he commented on the games in there. I mentioned that I needed to pick up the latest expansion for SOASE, he opened notepad and copied a serial key for the expansion and said there you go.

Oh, and the very next patch resolved my issue and I could play no problem.

Blogger dienw October 16, 2020 9:46 PM  

Vox, your post is a jewel to be written down and remembered by any creative person. I copied into my artist's journal Crawford's comments on tactical and strategic creativity.
His comments are a localized implementation of the book of proverbs principles of "perceive (pull apart mentally: Koestler's whole/holon), understand (how do the parts work and how do they work within the whole), and discretion(the plan and the act)."
Thank you.

Blogger Scott October 16, 2020 9:47 PM  

I would like to know that as well Marc. W. Miller is still around.

Blogger Unknownsailor October 16, 2020 10:06 PM  

Yossarian wrote:This is great. An old friend recently got in touch and asked me if I wanted to make a game. Then Nvidia releases its 3000 series. Then AMD announces their Ryzen 5000 series. Now this blog pops back up.

Things are moving so fast in this realm that soon enough hardware limitations will be a thing of the past. Devs will no longer have to worry about optimization and they will only have to focus on creation.

No, they won't. Unoptimized games are buggy, slow, and the people who actually play them hate them. You can't throw enough hardware at software that is garbage and get gold.

Blogger Snidely Whiplash October 16, 2020 10:52 PM  

Yossarian wrote:Things are moving so fast in this realm that soon enough hardware limitations will be a thing of the past. Devs will no longer have to worry about optimization and they will only have to focus on creation.
Ya know, I head the same thing when the Pentium was released, when Windows was ported to Alpha CPUs, when the Itanium was announced, with the introduction of 64-bit AMD CPU, with the release of the first 1GHz processor, and probably a few other times too.
There is no such thing as throwing enough hardware at a bad program to make up for bad programming.

Blogger Archella October 16, 2020 11:43 PM  

DevGame is just rocking now, so sweet.

Blogger dangerdog October 16, 2020 11:48 PM  

Brad Wardell was way ahead of other 4X game designers at the time. Gal Civ one and two offered so much customization for for the ships. You could design the entire fleet for you nation from the ground up. Also the civilization had so much personality and the "world" was so rich. That interview was great read, thank you.

Blogger Doktor Jeep October 16, 2020 11:56 PM  

After Mordor falls there will be a renaissance of video games technically and culturally.

Blogger Rick T October 17, 2020 12:02 AM  

The first law of programming is there is always one more bug. The second law is programs will expand to consume all available resources.

Exactly why does a copy of Chrome with 1 tab open need 11 threads and 580MB of RAM??

Blogger Snidely Whiplash October 17, 2020 12:25 AM  

@Rick T
Another law of programming is that you can always make the program run better and faster by removing a line of code.
So the ultimate program will be one line of code, running at temporal paradox speed, consuming the entire universe, but it will be a bug.

Blogger Valar Addemmis October 17, 2020 12:37 AM  

Yossarian wrote:
Things are moving so fast in this realm that soon enough hardware limitations will be a thing of the past. Devs will no longer have to worry about optimization and they will only have to focus on creation.


Not to pile on too much, but you'd benefit from the concepts at this page:

https://infogalactic.com/info/Wirth%27s_law

It's obvious to any person observing software development trends that we haven't really made "progress" as beefier platforms came out. Instead, software designers just beef their work up badly because the hardware can compensate. There's a rational case to be made against premature optimization, and that it's worth the performance hit because the coder time costs too much relative to the gain from optimizing on the front end instead of just at your bottlenecks. But it's futuristic singularity nonsense to believe that hardware will alleviate the need to care about doing things well.

Blogger Yossarian October 17, 2020 1:16 AM  

Snidely Whiplash wrote:Ya know, I head the same thing when the Pentium was released, when Windows was ported to Alpha CPUs, when the Itanium was announced, with the introduction of 64-bit AMD CPU, with the release of the first 1GHz processor, and probably a few other times too.

There is no such thing as throwing enough hardware at a bad program to make up for bad programming.


I was talking about polygon count, rigging, baking, ray tracing, and so on... but ok boomer, whatever you say.

Blogger Beloved October 17, 2020 2:01 AM  

@Snidely The program could be improved by removing the last line of code. Meaning the ultimate program doesn't exist.

Blogger Jack Amok October 17, 2020 2:17 AM  

Another law of programming is that you can always make the program run better and faster by removing a line of code.
So the ultimate program will be one line of code, running at temporal paradox speed, consuming the entire universe, but it will be a bug.


I think that's the same concept I try to teach youts getting into the profession, but I express it a little differently.

Complexity - not time or money - is the currency with which you pay for software. You start with a program that is incredibly simple to understand, but doesn't do anything. Then you start adding features. Every feature adds some amount of complexity, and eventually you've added so much complexity, the next change you make will mean you no longer understand the program. At that point, whether you want to or not, you're done.

The worse the programmer, the more complexity they have to pay to add a feature.

Blogger Yossarian October 17, 2020 2:23 AM  

Valar Addemmis wrote:It's obvious to any person observing software development trends that we haven't really made "progress" as beefier platforms came out.

I can pull off Pixar-level cloth simulation on a laptop that's several generations old using commercially available software. If there's a will, there's a way. Always, no matter what some wikipedia article says.

Blogger Jack Amok October 17, 2020 2:36 AM  

It's obvious to any person observing software development trends that we haven't really made "progress" as beefier platforms came out. Instead, software designers just beef their work up badly because the hardware can compensate.

Depends on the dev. Beyond some pretty simple gotchas, the majority of modern C/C++ devs are not going to do better than the compiler in optimizing their code. Not in the old-school sense of optimization anyway. The majority of sub-optimal loops in the 3GL language will be optimized into machine language by the compiler far better than they would be if the 3GL programmer tried to optimize his loop. In fact, by trying to be "clever", there's a good change he would prevent the compiler from doing the proper optimization.

Optimization matters at the cutting edge, and that moves. Optimizing a physics engine will make a far bigger difference today than optimizing a blitter. The people who made your compiler are better at optimizing your blitter code on the fly during a 30 second compile that you are with two months of contemplation.

Blogger Unknown October 17, 2020 7:53 AM  

VD wrote:What do you mean by "crashing totally off the rails"? I have some ideas about what you might be talking about, but I'm curious to hear your perspective.

The game went from an expanded set of rules, still playable with the basic rule set mechanics, into Mega Traveller. MT was more like an overly detailed war game plus accounting, not just for vehicle and ship operations but even for personal combat, and even task skills were difficult.

The decision to destroy the setting of a stable core Imperium with adventures at the huge and many fringes and outward from them, was thrown away for no good reason. Space vampire ships and instant tech reversions were weird rather than fun.

The attempts at computer games were poorly done and mostly unplayable.

Traveller 4 and 5 are bug ridden and over detailed messes. The attempt to license the setting to GURPS was well done, great background, lots of setting and materials, for those who chose to use that system. So of course the license was cancelled.

Mongoose Traveller seems like a Hail Mary play that stuck, so I would expect the license to be cancelled soon.

Blogger Mathias October 17, 2020 9:33 AM  

Yossarian,

It simply isn't true, because you are not getting near the performance increase you think you are with the upgrade. Yes, IPC is way higher, but the brute reality of rising CPU performance vs relatively stagnating RAM performance means good optimization in the form of good cache page access planning is more essential than ever. The primary culprit in sabotaging this is OOP, which is quite possibly the cache unfriendliest algorithm/programming pattern ever devised, with the exception of things like certain hash table algorithms. One of the big issues is that OOP was sold to the public at a time when the delta between memory access times and CPU performance was much smaller, and the design of OOP made way more sense. But then, over the years, as the clock cycle of the average CPU grew shorter and shorter, the physical gap between the RAM and CPU has become more and more relevant, as the propagation time of an electric signal is mostly fixed (you have to wait more cycles for a memory block to arrive as the cycles get shorter, since the RAM is not getting any closer to the CPU physically on a given motherboard). Of course, the designers of OOP languages have been trying to work around this by designing virtual machines which accept OOP on the outside, but convert it to something a bit more data driven on the inside, which is vastly more performant on modern hardware. This is a difficult and complex task however, and the results do not tend to be entirely backwards compatible with preexisting software.

Blogger VD October 17, 2020 9:36 AM  

Being a distant, but friendly acquaintance of Marc Miller, I think the problem is that Marc is a systems and process guy. He needs to be teamed up with storytellers in order to be most effective. Like Chris Roberts, he can achieve greatness with the right support, but tends to be less effective when it is lacking.

I not only disliked MegaTraveller myself, I completely ignored it. I tried to convince him to re-release the Little Black Books, but he was understandably focused on T5, which he had just finished, so he was not interested.

Blogger Valar Addemmis October 17, 2020 10:45 PM  

Yossarian wrote:Valar Addemmis wrote:It's obvious to any person observing software development trends that we haven't really made "progress" as beefier platforms came out.

I can pull off Pixar-level cloth simulation on a laptop that's several generations old using commercially available software. If there's a will, there's a way. Always, no matter what some wikipedia article says.


Once a problem has a closed form solution, it's trivial to repeatedly solve it? Tell me more, oh software genius.

Your claim here is very different than the claim I was responding to. No interest in trying to enlighten a dim wit. I see other folks were right to peg you as a waste of time.

Blogger Shimshon October 18, 2020 3:15 AM  

I remember Chris Crawford from the Atari 800 and De Re Atari, having owned both. A great book. The 800 was quite a design. There were some great games for it. I have fond memories of spending many a night playing M.U.L.E with my buddies in college. My understanding is that its ludicrously slow but technically interesting serial bus design was the model for the later USB.

Blogger LastRedoubt October 18, 2020 4:10 PM  

@23 Unknown

I cut my teeth on the big black book, and the CT CD-ROM is a treasure trove. Never found MT/etc. to be worth the money, did like what became Cepheus.

That said, I'm glad that a friend of mine dragged me back to T5 as of 5.10 (a number of my recent posts at thelastredoubt.com cover it)

You can dive very, very deep into the systems, and there are still some holes (take melee rules...).

BUT

Marc knows exactly why he made the baseline assumptions he did, and how and why they foster gameplay.

He explicitly and often states that a number of modules should not even be used unless dramatically or otherwise necessary. You don't need every detail of a star system - just the basics of the main world. Everyday equipment doesn't break every day - so QREBS isn't an issue unless it might become a dramatic lynchpin.

And in practice, the character generation rules are simpler than the mass of BS that D&D3.5/Pathfinder had become. What you put down on the character sheet is certainly far simpler to work with.



Blogger Yossarian October 18, 2020 9:04 PM  

Mathias wrote:Yossarian,

It simply isn't true, because you are not getting near the performance increase you think you are with the upgrade. Yes, IPC is way higher, but the brute reality of rising CPU performance vs relatively stagnating RAM performance means good optimization in the form of good cache page access planning is more essential than ever. The primary culprit in sabotaging this is OOP, which is quite possibly the cache unfriendliest algorithm/programming pattern ever devised, with the exception of things like certain hash table algorithms. One of the big issues is that OOP was sold to the public at a time when the delta between memory access times and CPU performance was much smaller, and the design of OOP made way more sense. But then, over the years, as the clock cycle of the average CPU grew shorter and shorter, the physical gap between the RAM and CPU has become more and more relevant, as the propagation time of an electric signal is mostly fixed (you have to wait more cycles for a memory block to arrive as the cycles get shorter, since the RAM is not getting any closer to the CPU physically on a given motherboard). Of course, the designers of OOP languages have been trying to work around this by designing virtual machines which accept OOP on the outside, but convert it to something a bit more data driven on the inside, which is vastly more performant on modern hardware. This is a difficult and complex task however, and the results do not tend to be entirely backwards compatible with preexisting software.


Graphics are offloaded to the GPU, which is why they were invented. I don't care about your RAM/CPU discrepancy talk. I doubt anyone here has even pushed a computer to a breaking point without using some benchmark. I was talking about a specific use, i.e. video games, and graphics are the most expensive and time consuming part of game development. That's slowly becoming not the case anymore.

Valar Addemmis wrote:Your claim here is very different than the claim I was responding to.

Yes. You created a strawman in your head, then attacked it with some epic wiki link, then got all mad for some reason when I made my original point clearer. It's obvious by the way you think of computers that's you're just a basic coder. I know you people, I know the way you think, I know that you can only follow instructions, and I know you think others care about what your opinion of them is.

Blogger Valar Addemmis October 19, 2020 11:56 AM  

Yossarian wrote:

Yes. You created a strawman in your head, then attacked it with some epic wiki link, then got all mad for some reason when I made my original point clearer. It's obvious by the way you think of computers that's you're just a basic coder. I know you people, I know the way you think, I know that you can only follow instructions, and I know you think others care about what your opinion of them is.


Replying to the words on a page instead of the words in your head is a strawman attack? Such charming and innovative argumentation. Funny that someone who seems to want to be able to just script "creativity" has the nerve to call someone a "basic coder." You know nothing, about me or any topic I've actually seen you participate in on VP. I'll remedy the mistake of engaging in the future.

Blogger Canadian Warlord October 19, 2020 4:53 PM  

I love hearing about the games I missed. Might be worth turning around and going back.

Post a Comment

Rules of the blog

<< Home

Newer Posts Older Posts