The Cost of Jaguar Development

[sh3-rg]

Where's the missing CV.PDF?

[+CyranoJ]

Zarch (also known under its ported name of Virus) is a computer game written by David Braben (better known as the co-author of Elite) in three months in 1987, for the release of the Acorn Archimedes computer.

 

I guess it depends on the skills of the coder.

 

Regardless of how much money you pour into it, you cannot expect to get the full 3d engine and a game for a croc-style game done within one year from a coder that never wrote a SW rasterizer in ASM. That's completely unrealistic.

[VladR]

Zarch (also known under its ported name of Virus) is a computer game written by David Braben (better known as the co-author of Elite) in three months in 1987, for the release of the Acorn Archimedes computer.

 

I guess it depends on the skills of the coder.

 

3 months actually sound like awfully long time for a simplistic pseudo-3D game like Virus. But I suppose those 3 months weren't full-time (e.g. ~50-60 hrs a week), unless he was just learning to work with basic 3D space.

Or maybe he was exploring other concepts at the same time, that you can't see in the game. Otherwise, there's really not much to do to work on such a game for full 3 months, unless you are just starting.

 

I certainly hope you aren't entertaining the idea of coparing Croc with Virus, since it's like comparing Wolfenstein with Quake. Quite a leap in technology.

- Virus has fixed camera, that makes plenty optimizations screaming for grabbing

- Virus does not have animated polygonal enemies. To write a fast component that handles animations and all transitions between animations is quite a lot of work. Easily more than whole code of Virus.

- Virus does not have collision detection. No, comparing the distance between two centers of BBox does not count as ColDet

- Virus does not have Free Movement along Y-Axis - e.g. you can jump UP/DOWN on various obstacles, which is far more complex in 3D than in 2D

- Virus does not have pathfinding with obstacle avoidance, its AI is basically 2D (as is whole game, gameplay-wise, albeit with the working illusion of 3D)

- Virus does not have a proper Level Design, thus its code, once written, is guaranteed to work across whole game. Good luck with that with free-roaming game like Croc

 

Just because Croc has a kids-style art direction, does not make the code any easier, than say - SkyHammer, I-War or other free-roaming action 3D games on Jag.

Actually, since plenty those games just feature the static 3D meshes, Croc as actually much more complex because of the 3D animations of characters.

 

 

 

Of course, if you can hire coders that actually have an experience with SW rasterizers written in ASM, then you can get the prototype whipped up in, say, 3 months.

 

Good luck trying to persuade those to quite their cushy jobs for a project on a quarter century dead platform

[sh3-rg]

Has anyone really been far even as decided to use even go want to do look more like?

 

 

Sheesh that Braben guy was such a lamer, eh? Screw him and his pathetic 3-month 60-hour-week starting-out simplistic efforts. Screaming for grabbing a clue the guy was. Good luck trying to find a point to the post you shall need.

 

[+Gemintronic]

I think we have to focus on funding tools that allow people to use their own free time to create more Jaguar games. Spending money on a single game engine for a single game isn't going to have an impact.

 

I'd rather invest in, say, developer hours making the Jaguar a target for ZGameEditor:

http://www.zgameeditor.org/

[VladR]

I think we have to focus on funding tools that allow people to use their own free time to create more Jaguar games. Spending money on a single game engine for a single game isn't going to have an impact.

 

I'd rather invest in, say, developer hours making the Jaguar a target for ZGameEditor:

http://www.zgameeditor.org/

 

Well, everyone has his own dreams, if they win a lottery. You can't be talking about The Greater Good, when you are not footing the bill

 

I, for example, would fund an expedition to cross Antarctic on foot. No relation to Jaguar, but hell lot more fun, and besides - it'd be my money

[+Gemintronic]

The point is no one here can fund a single game. They should be instead focusing on investing in technologies that would allow other people to - as you say - do what they want: make games. By investing in the means they eventually get the game they want through enabling others.

[VladR]

I hear ya, but if someone's dream is to fund the development of a single game - and thus become part of the Atari history - and they actually foot the bill, you can't do much about that.

 

I think this summer - when I plan to switch to VRBasic from my PC coding, we'll get a bit closer to that dream, when I iron out the few missing features (OP,Blitter,Audio,Input)

[NinSEGA]

Wait, I'm a bit confused here. Being a super amateur at game programming, please forgive my ignorance.

 

If source code to a game of another platform is already available, how much more work needs to be done for the porting/conversion (not sure which term is the correct one to use here) to the Jaguar?

 

You have devs like The Removers who managed to port the PC version (or was it Amiga?) of Another World to Jaguar. They had their project paid for in advance by doing pre-orders. Surely that is something that can be done by others like Paul?

[+remowilliams]

You have devs like The Removers who managed to port the PC version (or was it Amiga?) of Another World to Jaguar. They had their project paid for in advance by doing pre-orders.

 

 

Well if you define 'paid for in advance' as covering parts of their material expenses then probably yes. Seb and RGC weren't paid for doing any of the actual work.

[VladR]

Wait, I'm a bit confused here. Being a super amateur at game programming, please forgive my ignorance.

 

If source code to a game of another platform is already available, how much more work needs to be done for the porting/conversion (not sure which term is the correct one to use here) to the Jaguar?

 

You have devs like The Removers who managed to port the PC version (or was it Amiga?) of Another World to Jaguar. They had their project paid for in advance by doing pre-orders. Surely that is something that can be done by others like Paul?

That depends how close the two platforms are in terms of HW features/performance.

 

If it's a higher-level language like C, then porting becomes much easier, since only platform-specific code for rendering/input/audio has to be replaced.

 

Then there is the question of how compatible are art assets. Is the target resolution the same ? Then you can simply reuse the art assets without having to redo/resize them - which would mean a lot of changes in the code (should their dimensions change). But certainly nothing insurmountable (just annoying).

 

 

I don't know about Removers, but I suppose their effort wasn't financially backed by an investment ? E.g. it wasn't a job. If it's a work of love, then the question of money is irrelevant.

[swapd0]

Why you wanna spend about $50k into a game if you'll never have some revenues?

 

Imagine that you spend $50.000 and some coders make Tomb Raider just like the PSX version (imagine it...), how many carts you'll gona sell? Maybe 100, 200, even if you sell 1000 carts at $50 each, you're loosing money.

[+remowilliams]

But I don't want to imagine Tomb Raider on the Jag...

 

 

 

Oh no. I just did

 

 

 

[+Gemintronic]

Two different programmers can have vastly different styles. Different compilers have different needs. Different video libraries have different calls. Simply having the source for system X does not make life easier for development on system Y.

 

Not saying you don't know that NinSEGA. Just pointing out some of the reasons why C source code does not guarantee a fun time when porting.

[NinSEGA]

Two different programmers can have vastly different styles. Different compilers have different needs. Different video libraries have different calls. Simply having the source for system X does not make life easier for development on system Y.

 

Not saying you don't know that NinSEGA. Just pointing out some of the reasons why C source code does not guarantee a fun time when porting.

 

And isn't programming for the Jag a pain in the ballzack?

[+Gemintronic]

 

And isn't programming for the Jag a pain in the ballzack?

 

Not with the recent assembly based libraries available. For the longest time there has actually been a Live CD that boots up and already has everything you need to code installed and ready to go.

 

Even a derpy BASIC dweeb like me was able to set up a C dev environment in Win 7 64 bit and compile a "hello world". Mind you, I forgot everything about C a decade ago.

[sh3-rg]

You have devs like The Removers who managed to port the PC version (or was it Amiga?) of Another World to Jaguar.

 

Seb describes in detail how he went about creating the Jaguar version of Another World in JagChris' interview.

[+CyranoJ]

I certainly hope you aren't entertaining the fact that all programmers work at the same speed, that would be an amazingly simplistic and opinionated thing to do. Quite a leap in gibberish.

 

As for your other comment... wow... just... wow... words fail me. Tell me, can you see from so far up there?

 

 

I certainly hope you aren't entertaining the idea of coparing Croc with Virus, since it's like comparing Wolfenstein with Quake. Quite a leap in technology.

 

3 months actually sound like awfully long time for a simplistic pseudo-3D game like Virus. But I suppose those 3 months weren't full-time (e.g. ~50-60 hrs a week), unless he was just learning to work with basic 3D space. Or maybe he was exploring other concepts at the same time, that you can't see in the game. Otherwise, there's really not much to do to work on such a game for full 3 months, unless you are just starting.

 

To put this in perspective, aren't we at about 3 months for Klax now? And isn't PoP over a year? What about Hero? I suppose your time isn't full time either..... There isn't much work to do in such a game, after all, Hero runs on a 2600. Bearing in mind, those games are already developed, it's just a straightforward simplistic port, with modern tools for added speed aids. Imagine doing those in 1987....

Edited April 16, 2014 by CyranoJ

[JagChris]

Since we're talking theoretically. I guess you could have the team start combing through the source codes already available there and organizing them into libraries and working out the kinks with all that.

[+Gemintronic]

To save time I put OpenGL in main. The source is all little endian but that's a RISC I'm willing to take.

[GroovyBee]

3 months actually sound like awfully long time for a simplistic pseudo-3D game like Virus.

Really? I don't think so. Braben was using a BBC Micro (a lowly 8 bit 6502 micro at 2MHz) to develop the code and squirt it into the development system's RAM whilst targeting an ASIC that didn't have a fully complete instruction set (missing the multiply instruction).

 

As for your other comments on Zarch go, its pretty clear you haven't played that game either.

[+CyranoJ]

You could always JERRY rig it, but that'll be a job for TOMmorow.

 

To save time I put OpenGL in main. The source is all little endian but that's a RISC I'm willing to take.

[VladR]

Really? I don't think so. Braben was using a BBC Micro (a lowly 8 bit 6502 micro at 2MHz) to develop the code and squirt it into the development system's RAM whilst targeting an ASIC that didn't have a fully complete instruction set (missing the multiply instruction).As for your other comments on Zarch go, its pretty clear you haven't played that game either.

Yes, you bring up a very important point I have completely disregarded - and that is a toolset and speed of the build/run cycle on the machines of that era. Remembering the build cycle when using Atmas II on Atari 800- it was a pain and was terribly slow for bigger programs. Which was the reason I switched to machine code instead of regular 3-letter symbolics.

 

Ok. I was wrong about those 3 months, since I didn't take into account dev tools. Modern tools - like Notepad++, where you can switch between multiple files , fold functions and most importantly see up to 100 lines of code make you forget the days when you could see maybe 20 lines and scrolling took ages, let alone compiling. Splitting the source code into per compiled chunks since it never fit RAM was another massive slowdown.

 

I did, however play the game some time ago and finished few levels to get a good idea what the game is about. It's basically a playable tech-demo, but I always loved those.

[VladR]

I certainly hope you aren't entertaining the fact that all programmers work at the same speed, that would be an amazingly simplistic and opinionated thing to do. Quite a leap in gibberish.

 

As for your other comment... wow... just... wow... words fail me. Tell me, can you see from so far up there?

 

 

 

 

To put this in perspective, aren't we at about 3 months for Klax now? And isn't PoP over a year? What about Hero? I suppose your time isn't full time either..... There isn't much work to do in such a game, after all, Hero runs on a 2600. Bearing in mind, those games are already developed, it's just a straightforward simplistic port, with modern tools for added speed aids. Imagine doing those in 1987....

So far, I have spent about 3-4 weekends - usually just about a day each weekend, maybe a full 7 days on Klax, altogether.

And all that time went into engine and optimizations. Unless someone else can beat my 3d rasterizer in C, I am lead to believe it is fastest C rasterizer currently present on Jaguar, so I believe it was a time well spent since it proved that a common misconception that C rasterizers are terribly slow, is totally incorrect. Even you, with alleged decades of experience on Motorola said plenty times it would take seconds for each frame to be rendered in C. But not all programmers are born equal...

 

As for Hero. I said probably dozen times that Hero code base serves as playground for all 3d engine experiments and for testing out the library for VRBasic. Of course, if it was just a 2d remake, it would be finished, but I find That useless.

 

And PoP ? What about it ?

 

Besides, repeating myself for umpteenth time, till summer, I am working on two PC games and took completely undesired detour with the Klax...

[+CyranoJ]

Ah, so three months of your spare time isn't the same as three months of David Braben's spare time.

 

He can accomplish a full game, you can barely manage a tech demo. But hey, go ahead and belittle his work.

 

'till summer, I am working on two PC games'

 

You should easily be able to accomplish that, I mean nearly everything is done for you on the PC already with fully working libraries. And there is oodles of RAM. No need for you to type in direct machine code instead of 3 char symbols.

 

I get it. You were proclaiming your superiority again over everything, while actually showing nothing completed. All good, time to move on.

 

So far, I have spent about 3-4 weekends - usually just about a day each weekend, maybe a full 7 days on Klax, altogether.
And all that time went into engine and optimizations. Unless someone else can beat my 3d rasterizer in C, I am lead to believe it is fastest C rasterizer currently present on Jaguar, so I believe it was a time well spent since it proved that a common misconception that C rasterizers are terribly slow, is totally incorrect. Even you, with alleged decades of experience on Motorola said plenty times it would take seconds for each frame to be rendered in C. But not all programmers are born equal...

As for Hero. I said probably dozen times that Hero code base serves as playground for all 3d engine experiments and for testing out the library for VRBasic. Of course, if it was just a 2d remake, it would be finished, but I find That useless.

And PoP ? What about it ?

Besides, repeating myself for umpteenth time, till summer, I am working on two PC games and took completely undesired detour with the Klax...