This is the question we get asked most frequently, so we thought we’d answer it in some detail.
Firstly, many people seem to imagine that it’s some kind of trivial oversight, as if we are going to say “gosh, yes, we hadn’t thought about that”, and then fix it. We have thought about it, of course. We think that it would be great for the player – and the robots – to be able to enter buildings. We’ve spent a lot of time examining how they might do that, and concluded that, some open and ruined structures aside, they cannot.
There are two main solutions to being able to enter buildings:
- Skyrim-style instanced buildings, where the interior is essentially another space. Enter through the door and explore that space independently. This is clearly a compromise, and not ideal for an open world game. What happens to the robots chasing you? Do they just wait outside? Do you get an outdoor camera? Do they come inside? It rapidly moves away from the kind of game we were trying to produce, as well as producing a cascade of additional design issues.
- Buildings that are entered within the world space, as in Arma II or Stalker. These can be explore seamlessly and allow you to look out into the world through the windows and doors. This is the ideal solution, and they one we would have to go for.
So we immediately discounted the instanced buildings, because they are not part of the world. They don’t allow you to hide inside and peek outside at your antagonists. They would require interiors to built, either statically, which would increase our art time and budget, or procedurally, which would increase our art time and budget as well as programming time (proc gen solutions are extremely intensive for the programmer and testers). Doing it procedurally would mean increasing the generation time for any given world enormously, too, which makes doing it like this a drag for the player. Do you increase the generation time at the start of the game by several minutes? Or create a lag on entering any building? Clearly, neither is ideal for creating a streamlined experience. If you do go for static interiors, then you have to create a large number of them so that you don’t appear to be entering the same building all the time, which again increases our art time and cost, and implementation, and testing.
The points above about creating the building procedurally also count towards the second more ideal solution, with necessary increases to procedural generation complexity and art budget. There are other questions too: do the interiors then have furniture? Where is the loot? Do we create additional 3D objects for the items, or do we create secondary loot containers inside the buildings, rather than the doors as per the current system? Solutions like those in Arma 2 are great, but they add several large burdens to development, not least of which is cost.
If we did decide we had the time and budget to go with that second system – and let me be very clear when I say that we do not have the resources, we are a three-man team with a small Kickstarter budget, and I still had to work another job for the whole of development – there’s another problem, too, which is that we’d have to introduce another complex level of detail structuring system to deal with the increased geometry load from the villages. These areas are already very intensive, even with a bunch of low poly models, and when we’re throwing this many polygons at the screen we start to lose slower machines entirely. The performance hit would be significant, and the polycounts of the buildings would skyrocket. Given that this was supposed to be a lo-fi game about British countryside we’re suddenly spending huge amounts of resources – computational and developmental – on one problem that is solved by saying “you can’t enter the buildings”.
So that’s what we said.
Oh, and then there’s the issue of having to build every model in such a way as to allow the robots to enter the buildings – or redesigning pathfinding entirely. The system that procedurally places a navmesh onto the world can’t cope with interiors, so we’d be looking at yet another solution that needs to be designed and programmed. More time. More resources, and another hit to the game’s performance.
It can be done. Yes, it can be done – other games have done it. But we, Big Robot, the guys making Sir, You Are Being Hunted before the money runs out, cannot do it.
Of course there are structures you can still enter in Sir – ruins, pole barns, and so on – but the interiors of the intact, British-style houses we would need to sell the “British” ideal were simply out of reach. And this is really important: we could have removed buildings entirely. We could have made it so that there was just forest and countryside, but part of the high concept of the game was Britishness, and we couldn’t get that across in the way that we wanted without British villages. And granted, we do lose something by not solving this issue – the capacity to explore and hide inside buildings – but since the point of the game was running across open countryside, rather than poking about in urban spaces, we were happy with that concessions. Lots of games do building interiors and urban environments well, but we do open, thematically-coherent countryside well. That was our goal, and we totally nailed it.
Yeah, I am bound to be a bit out of sorts when addressing this problem, so I apologise for that. It’s somewhat exasperating for us to create a procedural generation engine unlike anything else ever made, to have made Herculean efforts to produce a first-person stealth game with a huge range of weapons and equipment, to implement an AI system more complex than what generally appears in first-person games, and still be told that we’re lacking because the player has to use the buildings as loot stashes rather than Wendy houses.
We’re gamers too, of course. We can see the ideal game you all have in your minds. We’d love to have been able to deliver it, but the part about entering buildings is missing.
I hope you can understand why.
Most importantly: once you play the game, that concern will drop away. The game works, and players get along with the systems we’ve created just fine. Trust us, and wait for August 19th.