ARMA3 - Brand new engine or backward compatibility?

[maddogx 13]

Disclaimer: For the sake of this poll, I have made several assumptions that may not be correct, for example that BIS will eventually create a successor to Arma2, and that a complete engine rewrite would break backward compatibility (addons and missions) with previous BIS games, as detailed below.

Basic premise: For years, people have been requesting certain features, for example:

• updated physics (ragdoll, PhysX etc.)
  
• a complete editor overhaul (unit customization, weapons loadout)
  
• mod/addon switching without restarting the game
  
• object oriented scripting support (for example Lua)
  
• DirectX 11 support
  
• more customizable soldiers (face shape, uniforms, body type etc.)
  
• better unit control (fluid stance, less "clunky")
  
• advanced weapon handling and interaction
  

The list could go on forever. Now, it is my point of view that implementing many of these requests would require exorbitant changes to the current engine. So if BIS really wanted to include them in an upcoming game, maybe a complete rewrite of the engine would be better, even though it would break compatibility with older content?

For a longer explanation on why I think a rewrite could break backward compatibility see below, but first the poll options in detail:

Poll Options

1. Complete rewrite: You think BIS should completely rewrite the engine for and disregard backward compatibility.
   
2. Keep it compatible: BIS should build upon their current engine, ensuring compatibility with older content and disregarding new features that might break it.
   
3. They should do both!: You disagree that a new engine and backward compatibility are mutually exclusive and think they could do both.
   

Longer explanation:

Using the experience that has been gathered over the past decade with the first three iterations of the Real Virtuality engine, I think BIS could create a real monster of a game engine - given enough time and resources. (Which is the main problem, I guess.)

The drawback to this would be compatibility. For example, people have been requesting the possibility to customize unit loadouts in the editor and save them right into the mission, avoiding the necessity of changing the loadouts via scripting. Unfortunately, the current mission .SQM files do not provide this kind of functionality.

Addons are another issue. Even between OFP, ArmA, Arma2 and Arrowhead there have been issues of compatibility, where models and configs required modifications in order to work properly. With a new engine that allows customization of units, deeper interaction with weapons and other features that I can't anticipate, I'm pretty sure that backward compatibility with older content would be impossible to maintain.

One example for requested features that - in my opinion - would require a complete new engine, is the matter of switching mods on the fly. It would be great if it were possible to simply disable or enable addons while in the game, so that joining multiplayer servers becomes less cumbersome. Imagine you have mods loaded that a server doesn't allow, but instead of simply blocking you from joining, it says:

"Addon X, Y and Z are not allowed on this server. Disable these addons to join? (Yes, Quit)"

or

"This server requires addons A and B, which you do not have. Download and activate to continue? (3.5MB - Yes, Quit)"

This is not currently possible, and whether it ever will be is doubtful - at least not without having to restart the game. It has been like this since the OFP days, and I'm assuming that the way mods are loaded is a core part of the game that cannot easily be changed. But with a new engine, designed from the ground up for this kind of functionality, surely it would be possible?

Anyway, these are just my opinions, so maybe someone who knows better can enlighten me. ;)

Edited July 31, 2010 by MadDogX

[DaveP 66]

Although it would be nice, I can't imagine BI would want to take the resource and time cost to do a total new engine game rewrite. There's always solutions to adding more to the game, it's just a matter of doing them. With possibly one or two exceptions I don't see how any of the features you mentioned couldn't be put into the next iteration of the engine

[namman2 0]

if they add physics i'll be happy (at least simple car physics )

[maddogx 13]

Although it would be nice, I can't imagine BI would want to take the resource and time cost to do a total new engine game rewrite. There's always solutions to adding more to the game, it's just a matter of doing them. With possibly one or two exceptions I don't see how any of the features you mentioned couldn't be put into the next iteration of the engine

Well, the phrase "complete rewrite" should perhaps be taken with a grain of salt, since no software developer will discard a perfectly good piece of source code. Many elements of the current engine could surely be reused, so it's not like they would have to start entirely from scratch.

You are of course correct though: most requested features could probably be added to the current engine, but you have to ask yourself how much work it would be to rewrite existing engine components to facilitate that. Sometimes the necessary effort is simply too much.

This may be a stupid analogy, but think of it like if you got into an accident with your car - repairing the damage might be possible, but if the costs of doing so are greater than buying a new car, then it's a total write-off. So you buy a new car instead. (Sorry for this comparison, BIS. :o)

What I'm basically saying is that at a certain point, implementing all these advanced features would require so much rewriting of the engine, that it may actually be "cheaper" to create a new one and design it that way from the ground up.

[DaveP 66]

Perhaps so; My reckoning is the work required to either adapt an existing engine to suit the needs of the game or make a new one entirely would be gargantuan. Look at the problems the 1944 D-Day game crew are having trying to put together a large scale realism based game engine

[nzdfcrash 33]

updated physics and a more fluid animation system for troops and animals would be enough to sell me on it

[maddogx 13]

Perhaps so; My reckoning is the work required to either adapt an existing engine to suit the needs of the game or make a new one entirely would be gargantuan. Look at the problems the 1944 D-Day game crew are having trying to put together a large scale realism based game engine

Yep, that's really the core of the problem, like I mentioned in the first post. Creating a new game engine is a huge amount of work, requiring lots of time and manpower that BIS apparently (and unfortunately) doesn't have. Being a relatively small company, they will presumably be sticking with what they have and slowly evolving their existing engine, as we have seen over the past ten years.

Unfortunately for us, this also means that it will most likely never quite reach the level of pure awesomeness that I think BIS would be capable of, if only they had more resources at their disposal. Imagine if some venture capitalist were to walk into the BIS head office and say:

"Hey Marek! Here's a billion euros. Now go make an engine that knocks everyones socks off. Take all the time you need."

:D

[hellfire257 3]

I'd certainly pay more for a new engine.

[Herr_kalashnikov 10]

A complete new engine would be the way forward. Incorporating all the knowledge and avoiding made mistakes. But if BIS hasn't started on making it I doubt we will see it within 5 years.

[CarlGustaffa 4]

Complete rewrite. However that needs funding. Where does the money come from? And I don't expect "the next game" with new engine be ready within the next 8-10 years either. So what happens in the meantime, does Arma die a slow death?

I propose getting out some DLCs to help funding (even if I'm against them), and have another OA'ish release in a few years containing mostly the DLCs and all units (OA and Arma2) on a new island and new campaign. But without grand changes to the current engine. Engine coders would be in pre engineering, brainstorming, and study phase now, side by side with the content makers who's working on something else.

Keep the content makers busy while the engine is rewritten, it will be a long time before a progressing new engine can be really utilized.

There is a tonne of features that all of use wants in new game, or new engine. OP listed some of them, but the list is extremely large. For many of these, engine changes are needed. For others, a complete rewrite is needed. But now take a look at what we already have and compare with other games. I fear we will loose basic simulation that we take for granted. Check out "that other game" who tried a "new" engine, not for what they have, but for what they lack. The game runs smooth, way better than Arma. But at what cost? When we get "all those new features in", do we really expect to have all existing ones, and getting the same performance? "That other game" doesn't even have fuel consumption supported; not a basic feature I would like being forced to script if the devs decided it would have to go to support new features. You can't add forever without any penalties.

The biggest danger of starting from scratch is how do you identify the mistakes? You don't have the amount of testers and you don't have the documentation. With RV they have a decent userbase who provide input and desires here and on the tracker. Designing from scratch, how do you even keep track of previous mistakes? And finally, what defines a "new engine"? Engine up with CryoEngineSomething or RealVirtousoEngine with a high resemblance to the old one. I know I'm pretty amazingly good at keeping old bugs and malpractices alive in new projects - bad habits die hard etc.

But there's more to an engine than an engine; most or all of the tools for using it might need a rewrite from scratch as well. Not to mention all the recyclables (content) that might have to be dumped; those configs doesn't write themselves, and those configs are anything but small.

Backward compatibility is thrown out the window (which it should), but I'm not sure about LUA. Where will LUA be in 10 years time? And what are the benefits over the RV scripting engine? The whole BIKI would be outdated instantly; now, how many years did it take the devs and all the users to put that together? Some "old timers" might fall off (I know I would - in ten years I'll be lucky if I can still speak let alone learn another language). I think a C++ API (similar to that in VBS, but I'm really out of my league here - not fully sure what I'm talking about) would be a better approach than object oriented scripting.

So the more I think about the sheer cost of rewriting the engine from scratch, the more I derail from my vote and end up more on the "keep updating it". Some parts can be fully replaced, but why reinvent the wheel if all you want is make the wheel more round and maybe add one?

[echo1 0]

In before someone suggests using whatever latest engine Crytek has in a non-ironic way.

*deploys parachute*

So the more I think about the sheer cost of rewriting the engine from scratch, the more I derail from my vote and end up more on the "keep updating it". Some parts can be fully replaced, but why reinvent the wheel if all you want is make the wheel more round and maybe add one?

Congratulations son, you understand the realities of software development... Now, if only everyone could follow suit...

Edited July 31, 2010 by echo1

[TechnoTerrorist303 10]

I stopped reading at the word "ragdoll".

Actually I don't think a complete re-write is needed but there are many potential features mentioned all over the forum that should be added. A better physics system is one of these.

Personally I hope that Arma 3 isn't for a long time yet and BIS use the time and the money generated from DLC to make it perfect.

[Inkompetent 0]

Even though it is a huge task I vote for complete rewrite. If they are going to write a new game engine, rather than just updating the current one here and there, then rather make it proper and drop any and all compatability.

In the end porting stuff (at least scripts) isn't generally a that huge deal almost regardless of language, so lots of functionality could fairly quickly be brought from ArmA2 to ArmA3.

For models I totally understand that a completely new engine probably will mean completely new rigging of the models and a lot of work, not to mention new fancy shaders and other technical stuff.

If their funds allows it a completely new, more advanced engine would be awesome. Stuff like ability to alter config-values on a per-instance-of-class-basis (for example modifying the lifeTime of fired shells/grenades), better integration of 3rd party applications, vastly modernized editor that can load, manage, and write scripts from within the game, in-game access to the RPT file, picture-in-picture rendering (MFD-screens) and clickable switches in vehicles, shooting with hand guns from vehicles, proper vehicle damage systems, and so on would all probably be a lot easier to include if rewriting the base rather than forcing them into the game with makeshift solutions because the very base of the engine can't manage the stuff properly.

I don't mind if BI writes a new engine now or later. But they have to at some point. And when they do, screw backwards compatability. Make it work. Make it good.

[CarlGustaffa 4]

In the end porting stuff (at least scripts) isn't generally a that huge deal almost regardless of language, so lots of functionality could fairly quickly be brought from ArmA2 to ArmA3.

How quick is fairly quickly?

How many man hours do you expect it takes to rewrite the 868 .fsm files and 2461 .sqf files I found for a selected few (I ran out of diskspace, lol) core Arma2 addons (includes mission files though)?

[jdmt 12]

I honestly don't care what they do, as long as it is realistic, and is the successor to ArmA II, I'm buying it.

Although, the following would be nice:

Interactive interiors for vehicles

Physics

Change weapons on the fly(Grab an M4, go to the ammobox, and grab an M68 CCO, and an M203, and attach them).

Render to Texture.

[Big Dawg KS 6]

if they add physics i'll be happy (at least simple car physics )

Are you implying that there are "no physics" in ArmA? FPDR

Also... Not. Another. ArmA 3. New engine. Thread. FPDR

You're missing a poll option: "The whole idea is completely absurd."

Edited July 31, 2010 by Big Dawg KS

[danny96 80]

Agree with completely rewrite. I wanna to customise my profile like in Rainbow six vegas 2 or ragdoll like in UT3. Nice suggestion!

[MoS 0]

It is totally useless to discuss a new engine for ArmA3 (if it is ever going to be created at all).

It just won´t ever happen and the reasons for this are simple: BIS is a type of family company, which could only survive until now, cause they found a gaming-niche, to which they created their own system.

They keep updating this engine for over 10 years now and it would be an impossible task to create a new system, which would be sufficient to meet modern requirements, like the acutal one does.

First it would need a huge amount of money let alone for the development, as they can´t just take an engine from another company and modify it to their pleasure. There is no comparable engine out there, which can perform the same way, with the huge landscape and the load of AI procedures, like the current one does, because as I already said, ArmA is a niche-product.

Second the whole BIS team would have to be trained to use the new technologies, what would also need

a huge amount of time, money and effort.

BIS can not afford to spend all this effort to make a game which then would also have to meet the expectations

of every ArmA fan out there and the gaming-scene.

To come to an end: an objective cost-benefit calculation would point out a high cost

with a low benefit (as this supernew game would also cost max. 50 euros).

The only thing we can hope for is that a business competitor picks up the idea and priorities of BIS

and makes their own new game from scratch (as seen in OFP: DR - the biggest fail in recent years)

[gossamersolid 155]

I don't want a complete re-write. As much as something like LUA would be great for modding, a lot of people don't know object-oriented programming. Therefore we'd probably have a lot less scripters and addon makers around for a while. BIScript is very easy to pick up on.

Plus people have to stop requesting ragdolls, it's getting annoying.

[Murklor 10]

Are you implying that there are "no physics" in ArmA?

There are "sort of" physics :rolleyes:

The handling of vehicles remain quite weak, but it works on a large scale I suppose.

Regarding the topic, I hope that backward compatability isnt going to hold BIS back... I mean really, with good enough tools how long would it take to "remake" a say a modded vehicle? You already got the base data, just need to export it with the proper rigging again.

PS: We need ragdoll!

[Prydain 1]

A complete revamp for animations, improved physics and adding a bit of fineness will suffice for me.

[DeclaredEvol 10]

Ragdoll, Realistic Explosions, Realistic Gun sounds and explosions like the HARCP and HiFi sounds, Realistic Radio Transmissions[ROFL], Realistic deaths[Get shot in the leg and you limp or get shot in the head and it knocks you're head half way off LOL], Explosions on players will rip off different parts[skeletal Meshes], Directx10 Support.

Btw, back in the day I was in the marines. I encountered tons and tons of different explosions along the way so let me help you guys with the sounds. Most explosions are extremely low frequency, as in it sounds lower than a bass nailing the E cord but more like a drummer kicking the bass drum. However the explosion creates a sound wave that travels 3 seconds late from about a distance of 200-300 meters. Maybe less actually, but when it explodes the sound wave will create a shock in you're body. Now the one thing people don't realize is that explosions are different on a cement, stone, wood, grass&dirt, and brick. When a explosions occurs on cement, it usually doesn't break the pavement but it causes a very powerful shock wave that can knock you down. Stone, explosions usually shatter the rock and make it fly pretty far... depending on the density of coarse. Wood, usually snaps and shatters making tons of huge splinters flying out at you're butt. Grass&Dirt usually explodes with grass&dirt flying up and pushing the everything else around the explosion radius outward. Basically flying with the shock wave. And at last finally brick\sandstone walls, always muffle the explosion but keep the smoke around. I don't really recall if it sends a shock wave if you're not exactly right beside the explosion. Any ways hope this shit helps you guys.

Edited July 31, 2010 by DeclaredEvol

[RaZoR1472 10]

Although i voted for a complete rewrite, this would only be in a dream situation (+ I don't want BIS to focus on making it backward compatible). A completely new engine from scratch is probably something that BIS don't really have the resources for. In all honestey, I would like to see BIS adding more VBS2 features to the commercial game. If they got all those features into RV2 I don't understand why they would leave them out or decide that we would not want these features for the general public. It's strange that despite being made with the same engine, the two feel and behave very differently (VBS2/ArmA1).

It seems to me that especially gun handling and effects are different. And, somehow seems more optimized? Maybe that was just the particular terrain that comes with JCOVE.

I would deefinitely sacrifice ArmA 2's graphical capabilities for more features and more "simulator" gameplay. Maybe if the campaign was a group of missions where you are an average soldier in a war zone, doing daily tasks that soldiers in the real world do. For example, maybe a a mission where you are a called upon to a local village where an IED has just gone of and some villagers are seriously injured. The team has to cordon off the area (laying barbed wire, redirecting traffic blah, blah. Call in immediate MEDEVAC from an an airfield and help load the casualties onto the Chinook (or whatever). Then the team calls for an observation team to go onto a highpoint whilst the playersr team seraches the surrounding area for enemy. Once the situation has been handled. The team goes back to their base where the mission ends. (getting carried away there:o).

Some more less obvious things could be AI medics. It is so frustrating when you have a casualty but the medic won't do anything. He will just continue to fire at the enemy position while his mate just bleeds out on the sand (which would be another cool feature;)). AI awareness is another big issue. It seams silly really to be able to fire at a group of enemy soldiers with out them reacting untill they have taken causualties. Just two AI based issues that need to be sorted.

I think something that would help BIS alot, would be to get some mlitary advisors onboard their team. This would help BIS to add the right features to the game. Instead of some of the stuff we see that we know BIS can do better with.

I understand that BIS may be looking at other things other than military "sims" but I would like to ask them to actually make a military simulator. Although this is the best thing on the market atm. I know that BIS can do better, they already did, 9 years ago:yay:.

In short, I would love to see BIS create a brand new engine from scratch with the best this and the best that. But I would rather they just simply focus on providing a military simulator first, Gucci and glamour second.

Edited July 31, 2010 by RaZoR1472

[STALKERGB 6]

Although its probably true that BIS could create a monster engine if given the funding and time I think realistically just reworking the current one (even drastically) is much more likely. Also in terms of compatability personally I probably wouldn't have the time or urge to re-learn how to mod in infantry for Arma 3, most modders may feel different but it requires alot of effort to get to even a decent understanding of making addons in arma (well for me anyway lol)

[Big Dawg KS 6]

There are "sort of" physics :rolleyes:

The handling of vehicles remain quite weak, but it works on a large scale I suppose.

Sort of? Let's see, we have:

-gravity

-collision

-forces

-acceleration

-momentum

-inertia

-etc..

Yep, all the basics of physics.

Honestly people, don't talk about things you don't understand. As for staying on topic, I won't even bother to try since the topic is nothing that hasn't already been discussed to death and will not have any different outcome this time.