Why Natural Selection 2 is not on Source Engine?

[Dmitriy]

Just curious why UWE chose to develop their own engine rather then use Source Engine for making Natural Selection 2?

Some reasons I see are:
- the need to pay royalties to Valve and for Havok physics
- it's good to have your own engine to be able to make new games using it or sell it to licensees

What are the others?

[Vitdom]

They had to be able to build the game as they wished with their own functionality. The source engine could not offer what Spark now has. Why pay to develop on an engine and waste time on for something you can't achieve with it in the first place? They were a start-up independent games development company and they were successful in creating a high-quality game.

Now after the release of NS2 they even have their own proprietary engine they can do whatever they want with or change in any way they want to make and add additional games to their portfolio; and they are already working hard on two new games, Subnautica and Future Perfect.

Devs explaining their thoughts about their engine in the most early days, answering your question:
https://www.youtube.com/watch?v=yH-SH1Iw6TY
Their main point in this video, the enhanced visual features etc are all engine-controlled in realtime. The source engine does not have this capability and every map has to be compiled where the compiler places lights, faces(walls) etc in the map.

If NS2 would have been made as a mod, then Source would have sufficed, but it is VERY dated engine technology and nothing you would consider as a start-up independent games company.

[Gotta Ramble]

Well, because source looks like sh.it is one great reason.

[Dmitriy]

So adding to reasons:
- It is faster to design levels with Spark Engine because of dynamic realtime lighting. While in Source Engine a map needs to be compiled for an hour
- Spark allows much more beautiful shaders(lighting) and effects. These shaders and effects couldn't be achieved with Source Engine(questionable).

[tengo chaleco]

because they cared more about lighting than performance and stability =gg

[Chris]

Frankly, I've been having more stability problems in CSGO than in NS2 recently

[hectoneon]

Originally posted by Vitdom:

Now after the release of NS2 they even have their own proprietary engine they can do whatever they want with or change in any way they want to make and add additional games to their portfolio; and they are already working hard on two new games, Subnautica and Future Perfect.

yet subnautica uses unity, make of that what you will. in my opinion developing an own engine was a mistake overall. yes, ns2 on source would have looked worse, but i think the horrible performance isn't even close to being worth the better looks. and who said the only options were source or an own engine? iirc even a dev stated on the uwe-forums that if he could do it all over, he wouldn't go for spark any more.

[Dmitriy]

did dev state why did he change his mind?

[Ghoul]

Ok, I'll shortly write down how is see things from my maybe a bit more insight vision:

Back in 2007 when UWE started to develop NS2 they made a prototype with Source and quickly realized that they could not make certain things they wanted to do with it.

When the first prototype of spark was shown with it's against source (at that time) superior lightning system they quickly choose to go with their own engine giving them the power to do whatever they want plus the potential of extra income by licensing the spark engine tools (best example is the decoda lua IDE)

Based on what i heard from the UWE devs the biggest thing they learned from developing NS2 is the following:

Creating game content while developing the engine is not a great thing to do. Basically often the engine team couldn't focus on optimization or things they wanted to do just because the game content team needed a certain feature to go on with their work.

That lead to the situation that a lot optimization work was just not done or was rescheduled to later dates.

Overall the spark engine is not as bad as many ppl declare it to be,actually it has some features which were ahead it’s time (like the dynamic lightning system)

The issue with it’s development was basically that the “engine” team at uwe just never got to focus on optimizing some parts or had to do some things the fast way just to keep up with the “game content creating” team.

This is one of the major reasons ns2 has performance issues and why spark 1 still has so much optimization potential even after 6 years of development.

(BTW looking at titles like the lately released AC:Unity UWE is not the only dev having to deal with this kind of issues.)

That said UWE is now not doing the same "mistake" again:

The game content team is currently working on Subnautica using a third-party engine which fit to their needs for a open-world-exploring game (unity).

Meanwhile the engine team is working on building a new and better version of spark taking the time to work at any part they want. This project and game is also allready avaible at steam as Future Perfect: http://store.steampowered.com/app/282370

Which is more a engine sandbox game like garry's mod (but uwe plans to add some decent gamemode on their own in the future before it goes out of Early Access).

So all in all the moral of the story is that you are always smarter afterwards and that UWE realized that they made some mistakes while developing ns2 but looking at their current projects you should realize that the certainly learned from their past.

Also building your own engine might be not always the best idea but overall gives you more freedom to do whatever you want to do. UWE did choose to chase after their dreams having in mind that their engine team is fully capable to create a powerful engine at their own.

Some solution UWE came up for Spark are also used at other games. So far at least i can fully understand their decisions (have in mind about that we talk about times were not that many third party engines did exist.)

For NS2 there is still imho a bright future because the CDT now has and takes the time to look at optimizing engine parts which never had been focused by the UWE devs before just as they did not have the time to do so.

[IronHorse]

Also it should be noted that at the time when they decided to make NS2, their only options for developing on pre existing game engines required hundreds of thousands of dollars to license. There was no UDK or Cryengine yet, and unity is only now starting to be robust enough for modern FPS.

So as an indie company (I think there was only 3 people on the team at the time?) they did not have funding to afford such a license. Combine that with the fact that Source didn't provide half the features they wanted and that Valve would not allow them to use their own tools in conunction with the Source engine license .. well, they were sort of stuck either making their own engine or not making the game at all.

[WelvynZPorter]

Originally posted by Hoxfort:

Originally posted by Vitdom:

Now after the release of NS2 they even have their own proprietary engine they can do whatever they want with or change in any way they want to make and add additional games to their portfolio; and they are already working hard on two new games, Subnautica and Future Perfect.

yet subnautica uses unity, make of that what you will. in my opinion developing an own engine was a mistake overall. yes, ns2 on source would have looked worse, but i think the horrible performance isn't even close to being worth the better looks. and who said the only options were source or an own engine? iirc even a dev stated on the uwe-forums that if he could do it all over, he wouldn't go for spark any more.

What horrible performance? lol

I play on the highest settings and never go under 60 fps.

[Naskoni]

Originally posted by Zii:

Originally posted by Hoxfort:

yet subnautica uses unity, make of that what you will. in my opinion developing an own engine was a mistake overall. yes, ns2 on source would have looked worse, but i think the horrible performance isn't even close to being worth the better looks. and who said the only options were source or an own engine? iirc even a dev stated on the uwe-forums that if he could do it all over, he wouldn't go for spark any more.

What horrible performance? lol

I play on the highest settings and never go under 60 fps.

So do I... on a i7-5930 and an nVidia 980GTX. Doesn't mean the engine is light and runs well on lower-end machines.

And yes, the system requirements are bloody worth the looks. The lighting in this game makes the levels for me. The game wouldn't run any better on Source though, not with those looks and not with those maps. Not if Titanfall is any indication.

[Dmitriy]

IronHorse, do you mean that Valve have forbidden them to use Hammer level editor and command line tools like VTEX, etc to compile textures, models, maps?

[100% Recycled Awesome]

Originally posted by Zii:

I play on the highest settings and never go under 60 fps.

Most people do not have the funds to buy the hardware necessary to run the game even at medium quality at 60 FPS. If you read the forums, both here and on UWE's website, it's pretty clear where the bell curve of hardware is placed and it's not in the higher end.