Moving TDM to new engine?

[NeonsStyle]

Is there any discussion, or thoughts from the devs, of porting TDM over to Unreal 4 or Unity? It'd be wicked to have TDM in Unreal 4, as it opens up so many

new ideas for level design, and of course much larger levels.

[Springheel]

Porting a game to a new engine is nearly the same amount of work as starting from scratch. A massive project like that is simply not realistic.

[kano]

Porting a game to a new engine is nearly the same amount of work as starting from scratch. A massive project like that is simply not realistic.

 

Even if it could be done over night, for free, I'd rather TDM stick with Doom 3 anyway than switch to yet another proprietary engine that will never be open source. The people who made TDM made all the right decisions. In addition to avoiding corporate politics, an open source engine ensures that the game can be pushed as far as the community desires. This is running on an engine from 1993.

Edited August 18, 2018 by kano

[HMart]

Is there any discussion, or thoughts from the devs, of porting TDM over to Unreal 4 or Unity? It'd be wicked to have TDM in Unreal 4, as it opens up so many

new ideas for level design, and of course much larger levels.

 

I'm not from the team but i know this has been discussed many times before and the consensus is, TDM will stay on the current engine for the foreseeable future.

 

They will not even port TDM to the idtech4 BFG engine that is just a evolution of the TDM engine, imagine porting to a totally different one like UE4, like Springheel said you would need to start almost everything from scratch, the current AI/actors/player controller can't be ported to UE4, for example, without major refactoring, all scripts need to be totally remade with UE4 blueprints, pretty much all missions made on DR would be rendered obsolete because they can't be ported to the UE4 editor, this would make all past and new mission makers really mad and would pretty much destroy the game.

.

BTW TDM would be impossible in Unity, not because of the game itself but because, imo people would be unable to make missions to it, Unity is a closed ecosystem, i'm not referring to closed source (that it also is) but that the way you make games in Unity, is totally different to the way you make games in idtech 4 and even Unreal Engine, on these the engine is a separate entity from the tools, in unity the tools ARE the engine, so if people would want to make missions for a Unity TDM they would need to have access to the full assets sources and the original Unity project and pretty much export a full TDM everytime they wanted to make a new mission to it. Imo that's the reason that pretty much all Unity games i know don't support modding.

Edited August 18, 2018 by HMart

[revelator]

Yeah moving to another engine would be a massive undertaking, and we could pretty much Kiss Any advances we made in idtech4 goodbye. Just My two cents.

[peter_spy]

this would make all past and new mission makers really mad and would pretty much destroy the game.

 

Switching to UE4 would have some significant advantages. Great performance boost, DX11 features, much, much higher performance limits, both static and dynamic lighting, set of modern, easy to use editing tools, etc. All these great things that just a couple of people working in their spare time would never be able to design in reasonable time. That is something that might help tremendously in getting new authors. But at the same time, current authors would be alienated, those who still rely on brushwork to create their maps. BSP support in UE4 is pretty basic, and there's no smoothing, as it's assumed that everyone uses models for everything these days. And that is in addition to what Springheel mentioned, so...

[New Horizon]

Is there any discussion, or thoughts from the devs, of porting TDM over to Unreal 4 or Unity? It'd be wicked to have TDM in Unreal 4, as it opens up so many

new ideas for level design, and of course much larger levels

 

Porting to a completely different engine is not a simple task. You can't just take code from TDM, drop it into Unity or Unreal and have it work out of the box. It would be years of work, essentially starting over and when it was done, you couldn't play any of the existing missions in it. From what I recall, Unreal and Unity are subtractive based engines, so the D3 additive built missions would just be left behind. It wouldn't be TDM, it would be a brand new Thief inspired modding platform. The time for porting to anything other than what we have would have been within the first year of TDM development, so probably back in 2005.

 

So, no. There is no discussion.

[peter_spy]

Unreal switched to additive geometry years ago, around UE3, and there's no need to build anything in sealed rooms anymore. You just place models in empty space and the engine uses LODs, distance culling, and other techniques for optimisation. But that's just a tidbit

Edited August 19, 2018 by Judith

[New Horizon]

Unreal switched to additive geometry years ago, around UE3, and there's no need to build anything in sealed rooms anymore. You just place models in empty space and the engine uses LODs, distance culling, and other techniques for optimisation. But that's just a tidbit

Ahh, I had forgotten it switched. Regardless, switching engines is still a moot point at this stage.

[peter_spy]

Absolutely. It would have to be some kind of TDM 2 initiative and making everything from the ground up, which doesn't make much sense now.

[HMart]

Switching to UE4 would have some significant advantages. Great performance boost, DX11 features, much, much higher performance limits, both static and dynamic lighting, set of modern, easy to use editing tools, etc. All these great things that just a couple of people working in their spare time would never be able to design in reasonable time. That is something that might help tremendously in getting new authors. But at the same time, current authors would be alienated, those who still rely on brushwork to create their maps. BSP support in UE4 is pretty basic, and there's no smoothing, as it's assumed that everyone uses models for everything these days. And that is in addition to what Springheel mentioned, so...

 

Sorry for the offtopic guys.

 

Yes there would be many advantages, there's no doubt UE4 is a fantastic engine and would open many new possibilities but in the case of TDM many important desavantages are true as well, you are forgetting the 100+ missions that are already made, all the stuff that could be transferred into UE4 would be, the mesh's, all the brush work, sounds and textures, all the rest, would need to be redone from scratch, AI/character system, player control, weapons system, scripting, lighting, particles and so on, like i said before, this would pretty much render those missions obsolete. Is not hard to imagine that many mission makers, would not be happy to scratch, the mission/missions they worked so hard on, also all the time they spent learning DR, how to make missions for idtech4 TDM, etc and start from essentially zero on UE4. I bet some T1 and T2 mission makers have no made the jump to TDM because of similar reasons.

UE4 TDM, would also take years to make, have no doubt about that, on that time, would mission makers just full stop making missions for current TDM and wait? If not, would the TDM team just trash all the missions made till the time UE4 TDM is ready? It could potentially be 200+ missions. Would the TDM team be courageous enough, to throw into the trash all the hard work, that all the past and present TDM engine coders did and are doing for the idtech4 engine and DR? Greebo works alone, for years in DR, focused on TDM, would they just forget all the contributions he did and say, thanks man but we don't need your tool anymore.

 

This may sound harsh but is real questions and stuff the TDM team need to weight before making so important decision.

 

Absolutely. It would have to be some kind of TDM 2 initiative and making everything from the ground up, which doesn't make much sense now.

This pretty much rendered my text above obsolete has well. ;P

Edited August 19, 2018 by HMart

[NeonsStyle]

Porting a game to a new engine is nearly the same amount of work as starting from scratch. A massive project like that is simply not realistic.

 

I know how difficult it is. I was just curious if anyone is thinking about it.

[peter_spy]

Sorry for the offtopic guys.

 

Yeah, I think this discussion is interesting, but our post should probably be moved to another thread, sorry.

 

One more thing in regards your last post Hmart – I think your depiction is a bit too dramatic, because of one simple thing: these projects would be treated as separate, just like making maps for e.g. UT2004 and UT3. Noone needs to abandon anything or transfer anything from TDM, as TDM 2 would be a completely new thing. Current version of TDM is advanced enough for current mappers to make stuff for another 10 years or more. TDM 2.05 already felt like complete package. 2.06 added many things, but also broke many things. So, in a few versions forward, with all of that fixed, you'll be getting the most ultimatest and bestest thing you could squeeze out of idtech4 (probably).

 

From my point of view, it would be interesting to see which mappers would come to UE4-based version of TDM, and how their maps would play out. Secretly hoping for AA or "indie AAA" quality

[wesp5]

So, in a few versions forward, with all of that fixed, you'll be getting the most ultimatest and bestest thing you could squeeze out of idtech4 (probably).

Isn't one of the reasons why TDM uses idtech4 that is has the best real-time lighting and shadows? I think later idtech engines went backwards and I can't remember ever seeing an Unreal engine game really showing convincing stuff in that direction. But maybe this changed with Unreal 4...

[peter_spy]

Unreal went with lightmaps, and added global illumination with UDK. In UE4 you have 2 types light rendering, forward (real-time) and deferred (baked). And, you can use both at the same time: https://youtu.be/qu3NfoXNZG4?t=4032

Edited August 20, 2018 by Judith

[revelator]

idtech4 used carmacks reverse, and while it is one of the best ways to do shadow volumes,

it actually had to be removed from the idtech4 source code because there was some debacle about if the patent belonged to creative (yes those guys that do soundcards of all things).

 

It was later hacked back in, by various people working on this engine, and also by us in some form, but not in its original form since that would probably net us a copyright infringment issue.

 

Portal culling also had a major impact on shadows.

[OrbWeaver]

 

I know how difficult it is. I was just curious if anyone is thinking about it.

 

Sure, people are thinking about it, in the same way that people are thinking about building a city on the Moon.

 

It's an interesting what-if scenario, but it's not going to happen in the foreseeable future.

[New Horizon]

Given the current state of the Thief modding community, i don't believe there are enough active content makers to support a split like that. It would literally have to be one platform or the other. Maybe if Eidos were to come along with a Thief game that resonated with the community the way the first two did, well...that might give it a shot in the arm but I think people have just moved on. I think TDM is doing quite well to update idtech 4.

[peter_spy]

idtech4 used carmacks reverse, and while it is one of the best ways to do shadow volumes,

 

Actually it's not. When you want to delete faces nobody will ever see, e.g. inside a cabinet, or you want to work with plane walls modules instead of boxes, this becomes a pain.

[nbohr1more]

Well, let's try to be honest about the current state of things.

 

UE4 and Crytech are now both Open Source and I believe that both now support a large number of

assets including most Id Tech 4 asset formats.

 

Someone made a TDM style lightgem in UE4:

 

http://forums.thedarkmod.com/topic/19457-pretty-neat-some-did-octahedron-darkmod-style-lightgem-capture-in-unreal-engine/

 

So on the surface, it looks like it would be easier now than back in 2009 when folks were trying to evangelize a move to Valve HL2 Source engine (etc).

 

So what's left after the above?

 

Biggest issue: AI code. Our Id Tech 4 GPL license is almost certainly incompatible with the license for UE4.

(Pathfinding, Stim Response, Weapon Damage, Door Handling, etc.)

 

If there is someway to reconcile the GPL licenses to allow our AI code to be integrated into UE4 then there is

some sliver of hope. If not the meat of the project would need to be rebuilt from scratch (no hope).

 

Entity Defs. We would need to either covert all Entities to whatever the equivalent is in UE4 or make some sort of

alias structure that reads id Tech 4 Entities and relays them as UE4 versions.

 

(See above AI \ GPL issue. If id Tech 4 code can be ported into UE4 then this can be made easier.)

 

Lights: We use projection images on lights. I'm not sure how much support for those exist in all UE4 render modes.

 

Particles: Id Tech 4 uses an exotic particle def system. We'd need to rebuild the equivalent effects using UE4 frameworks

 

Map format: We'd need to run maps through a converter as I recall. I don't believe UE4 has native support for Doom3 \ Quake 3 map formats.

 

 

Bottom line:

 

If the GPL code integration issues are resolvable, a good coder or two could get the meat of the project ported to UE4 in about 1.5 to 2 years.

 

If there is no legal way to merge in our AI code or other id Tech 4 (or derived code) then it would be a lot like starting from scratch. 4 years or more

of work.

 

With the limited number of contributors and time the core team has to work on this project, I will suggest that this won't happen from anyone here

regardless of which scenario is at hand.

 

An outside coder? Maybe. Maybe even that nifty guy who copied our Lightgem design for UE4? Who knows?

[chakkman]

You know, i would love that, but, the way it is now, with the engine TDM uses, it's fine with me. Actually, what i'd be massively interested in would be a Deus Ex like game like The Dark Mod, with custom missions, enhancing the orignal Deus Ex' gameplay. That in the Unreal Engine would be like WOW. Unfortunately, there seem to be few people as motivated as the TDM team, which is fair enough, of course, after all, it's something you do in your spare time, as a hobby.

[peter_spy]

FYI, converting maps to UE4 doesn't make much sense, as units are completely different now (1 uu is 1 cm, default character is 180 uu). Besides, models have completely different workflow (you need 2 sets of unwrapped uvs, one for textures, one for lightmaps, etc.). Remake is the only option. AI code and entities would have to be rewritten to match UE4 code and Actor Class structure. Lights have their own types and materials, no need for textures from TDM here.

 

IMO the biggest advantage here would be not as much to the audience (although PBR shaders and global illumination will significantly contribute to eye candy), as to content creators, at least those in the modern sense, so 3D generalists used to making their own models, textures, materials, particles etc. Everything is well-documented, and content creation is much faster, as is problem solving, since even coding can be UI-based. You don't have to create a huge wiki, code editor UI, or renderer, as there are people at Epic doing that for you. You can focus on content, gameplay, and game features.

Edited August 20, 2018 by Judith

[kano]

It would be more productive to get donations and then hire someone to upgrade our engine. That way we don't break everything the community has worked on, and we don't end up using features exclusive to Windows, or stuck with requiring a service like Steam/Origin, something that many proprietary game frameworks now require.

[HMart]

Yeah, I think this discussion is interesting, but our post should probably be moved to another thread, sorry.

 

One more thing in regards your last post Hmart – I think your depiction is a bit too dramatic, because of one simple thing: these projects would be treated as separate, just like making maps for e.g. UT2004 and UT3. Noone needs to abandon anything or transfer anything from TDM, as TDM 2 would be a completely new thing. Current version of TDM is advanced enough for current mappers to make stuff for another 10 years or more. TDM 2.05 already felt like complete package. 2.06 added many things, but also broke many things. So, in a few versions forward, with all of that fixed, you'll be getting the most ultimatest and bestest thing you could squeeze out of idtech4 (probably).

 

From my point of view, it would be interesting to see which mappers would come to UE4-based version of TDM, and how their maps would play out. Secretly hoping for AA or "indie AAA" quality

 

Yes i'm a dramatic guy but that is because i'm also a realistic, and i tend to measure well the pros and cons of everything i do, but imo i do need to be dramatic about this subject, why, because IMO the existence of two TDM's is totally impractical and divides the community in a totally incompatible way, so I believe the most realistic course of action, would be to abandon one of them, because unless there's two TDM teams, i'm not one for talking for the TDM team but my experience tells me, that the current one wouldn't want to constantly divide attention between this one and another, i sure working on the current TDM is hard enough.

Also imo people put to much hope into UE4, yes is a fantastic engine, yes the tools are superior, yes is "open source" but afaik, is not GPL compatible like idtech 4, you cannot take UE4 and make a commercial game and not pay a cent to Epic Games, ok technically you can, but for that to be true your game needs to be a flop and not gain much, not that TDM will ever be a commercial game but i hope you see my point, also idtech 4 is here to stay, idsoftware cannot remove it from the internet, because of all the different branches that exist now outside of their control, that makes the future of TDM totally safe, at lest has long the community stays working along. On the other hand, UE4 is only one engine, is in total control of Epic, they can remove the access to the engine at any time, so the future fate of TDM would be totally linked to what Epic does, not what the TDM community itself does.

 

Actually it's not. When you want to delete faces nobody will ever see, e.g. inside a cabinet, or you want to work with plane walls modules instead of boxes, this becomes a pain.

 

I don't remember that ever being a problem in idtech 4, i don't think it's stencil shadow system requires closed volumes? Afaik you can have flat planes cast shadows just fine. When i played with the dead C4 engine tho, it did required closed volumes, that until Eric introduced shadow maps.

 

You know, i would love that, but, the way it is now, with the engine TDM uses, it's fine with me. Actually, what i'd be massively interested in would be a Deus Ex like game like The Dark Mod, with custom missions, enhancing the orignal Deus Ex' gameplay. That in the Unreal Engine would be like WOW. Unfortunately, there seem to be few people as motivated as the TDM team, which is fair enough, of course, after all, it's something you do in your spare time, as a hobby.

 

Yes a motivated team like the one working on TDM is very rare indeed, it and SureAI are one of the best teams in the modding community that i know of.

[Obsttorte]

 

I know how difficult it is. I was just curious if anyone is thinking about it.

Unlikely, because if you would really had an idea of how much work it would be, you wouldn't ask

 

Besides that, what would be the benefit?

• If you could just port existing missions to UE4 for example (which you can't, but let's just image you could), they would not look (much) better in UE4. The same meshes, shaders and textures ... where should the benefit come from.
• Better performance? Maybe, as modern engines utilizes modern techniques that can help there. But still the way the engine works has to be taken into consideration, and the past has shown that most mappers won't do so anyways, so it would be just a bit less worse.
• Better ai? Same code, so no.
• Faster mapping? Theoretically yes, as UE4 doesn't use visportals and as Judith wrote sealing is not required (I don't know, so I just assume this is the case). But does this make things better? And is the gain that high.

Let's not forget that modern engines are heavely relying on using models for world building as mentioned above. Most mappers however utilize brushes and patches for building their missions. Something that was rather normal back in the day idTech 4 was designed, but is simple not state of the art anymore and will therefore most probably cause issues.

 

The most important part is that a lot of the work that would need to be done is coding work, and we only have a few programmers and tons of people that won't even invest their time in learning how to write scripts, besides them getting into "real" programming (it is basically the same, just more advanced). So who is doing all the work?

 

IdTech 4 is open source and therefore we can make all the changes we want. Soft shadows have already been included and there has been attempts on getting idTech 4 to utilize multiple cpu cores. Modern shader support has been added (unsure to which extent right now) and other stuff can be added to the code. This is much less work but could lead to the same results, so why should one take the complicated route of switching the engine which would in addition cause all the work that have been done already (or most of it) to become a big waste of time?

 

Btw: This has been asked in the past and all the answers given here have been given in the past. Doing a forum search (or utilizing google for that purpose) would have shown you. :|