This is a round up of the various activities related to the Nth Society that I'm aware of.

The project was publicly launched on Thursday 25th January. It has the mission statement:

To design, develop and play a game which explores life in a voluntary society.

The proposal laid out in detail the first conception of a game. In summary, it's to be a immersive, multiplayer online game with full ability for crafting, social interaction, trading, exploring, and all the things you'd want to be able to do in a comprehensive sim.

We received a lot of positive feedback. Started by @the-ego-is-you and I, we've been joined by new contributors which have joined us in a Slack chat for the last week, debating and discussing the ideas. You can join us too using this link.

Many of the discussions have centered around these topics:

1. How detailed and realistic should the simulation be?
2. How can we prevent bots from masquerading as human players?
3. Which game engine should we use?
4. How can a blockchain be used for a real time game?

1 - How detailed and realistic should the simulation be?

The main purpose of the game is to facilitate the exploration of real world social dynamics in a sandboxed world. So the question is how much detail is needed to make sure the simulation is "good enough"?

For example, do we need to allow for violence? And sex? Do characters have gender, race and other physical and hereditary characteristics?

There were a lot of different perspectives, but I think there was a consensus that some of these things are definitely need, and beyond the "basic set" we can leave which ones up to the game world creators. There is no decision yet on what constitutes the basic set though.

Anyone can create a game world, it's like starting a new blockchain from a new genesis block. Anyone who wants to play that game syncs to that game and can join in. It is proposed that the person who starts the game has some options which can only be set at that point, such as the fees for life creation, and could also include which "modules" to allow, such as sex or race of players. There was also discussion of making these updatable at a later time using player voting, but that idea needs development.

This allows many different simulations to run, for people to choose which one to join and not be exposed to or distracted by anything they strongly disagree with or are not interested in exploring.

2 - How can we prevent bots from masquerading as human players?

The game engine is based on a decentralized model, no one person "owns" the game. The game creator, the person who first started the game, has their special role end as soon as the genesis block is mined.

However it is (probably) desirable to only allow human players to play the game, to keep the simulation between people, i.e. a human society. But it should be noted that some contributors were open to the idea of bots performing a useful and interesting role within the game. I myself am unsure, but it's my perspective that human interaction is primary importance.

For much of the last week we discussed the idea of "challenges" being posed to players, when first playing and probably periodically (say once a day or once a week). They are basically like Captchas. These challenges should be obvious and easily solvable to humans, but their solution unclear to automatic systems. The best ideas were:

• Winograd Schema Challenge, which poses logically inconsistent choices which are however obvious to humans.
• Context determination of composite photographs

Winograd Schema

Here's an example, can you chose the correction answer?

The trophy would not fit in the brown suitcase because it was too big. What was too big?
Answer 0: the trophy
Answer 1: the suitcase

To chose the correct answer you must "identify the antecedent of an ambiguous pronoun in a statement" which is hard for an AI because "the task requires the use of knowledge and commonsense reasoning." [source]

This is good because they are easy to make, we use this kind of ambiguity all the time in speech and get along just fine.

It should be noted there is an active community of AI writers trying to solve this problem, but this is the case for any human verification method.

Context determination of composite photographs

This is a really novel approach, and there's a lot of ways to arrange this, but the idea is to provide one or more images and to pick the answer to a question.

Here's one I'm just making up now (Nth Society guys, I hope I'm representing this correctly 😅)

Image 1

Attribute: the Huffington Post

Image 2

Attribution: https://w-dog.net/ wallpapers

Question (pick one)

Who is having a great day?

1. Image 1

2. Image 2

You see, there's a kind of bait and switch you can do to make this enjoyable for the person taking the challenge, and also make the answer harder to search using Google Image search or similar method.

Other forms can include guessing the event that happened just before or that will happen just after the moment captured in a photograph. An example is answering whether or not a picture that shows that there is imminent harm.

Structure of challenges

We talked a lot about how to present the challenges, who should do it, should they be rewarded for the job and if so how, how to prevent cheating, how often to present challenges, etc etc and this discussion is very much still going on.

One thing seems likely though, that it will be optionally inclusive when creating a new game world, like some other features mentioned above. Perhaps you don't care that much about a few bots, or perhaps you really really do. That should be up to players.

3 - Which game engine should we use?

Discussions focused around MineTest, which is a MineCraft clone. When the proposal was created I wasn't actually aware of the existence of MineTest, so it was really cool to find out about it.

Attribution: MineTest.net

As an engine it's very attractive because it's in very active development, and really geared towards customization already. The game server is build into the same codebase as the client, making it perfect from the standpoint of combining both.

However it definitely has a more basic look and feel than some people were hoping for. While we could work with 3D modellers to replace the extremely simple objects with more detailed ones, some of the blockiness is of course fundamental to MineCraft/Test, so this needs more discussion.

Other alternatives were discussed such as Pixel Dungeon, especially since it is a mobile game. But this part of the discussion is still in early days.

4 - How can a blockchain be used for a real time game?

I mentioned before that we should no just "put a blockchain on it" to attempt to solve everything with blockchain, it's not a panacea. But since we need to track a distributed game state and make sure of the consistency of that without centralized authority, it makes sense to consider a blockchain for storing and modifying the game state.

The downside of using a blockchain is that it will be difficult to get the kind of speed or throughput that a real time game needs. This gets worse as the game scales up in players. The Steem blockchain is very fast, with a block witnessed / mined every 3 seconds, and each block containing multiple transactions, but even this would be limiting.

No concrete solutions have been proposed, but we have been exploring how operations might be made more efficient to process. Some ideas are

1. Use zoning to suspend processing of geographical zones which are absent of players. Requires some smart use of pseudorandom number generation to keep processing low but the future unknown.
2. Split players off from main chain while interacting in a zone, periodically updating and validating the main chain. This means high throughput between players in close proximity but they are not always validating the transactions of players which are far away.

It's still very unclear whether blockchain game state storage is viable, and what the alternatives are, so this needs lot of additional research.

Other activity

I wrote a post taking about item design, and especially the interaction between designers, artists and coders. It was a little technical, but there was a necessary level of depth to explain the subject.

In particular we need the items to be used in the game and their relationships to be mapped out. I'll put out a formal request for this soon, once we have settled on a first idea for a game engine I think.

Logo request with bounty

Every project needs a logo and Nth Society is no different! I put out a request for a logo and later added bounty for it. There's a total of 15 SBD up for grabs, it is still open.

We've had some great submissions so far and I'm super happy to see some interest from the #slothicorn community on it.

Disclaimer and final note

This is my representation of the discussions, in which I tried to be fair and balanced, but always note it is from my perspective. I do not speak for other people, I'm just trying to give you a window into the process. Others are welcome to make public notes on the process too.

Thanks for reading and hope you support and / or get involved in the project!