@SystemParadox I believe you are correct for a system based on locking. You don't have to use locks though. There are 100% safe lockless systems.

With the caveat that I've not actually implemented it, here is my (what I believe to be) 100% safe system:

1. Partition the segement into sub-segments. Each shard owns a sub-segment.
2. If we want to make a modification, read the entire segment, modify our sub-segment, write to the segment.
3. On future ticks we know what our sub-segment should look like. Read the sub-segment and check. If it doens't look right that means we have a condented write which the other guy won. Skip a random number of ticks (or wallclock time) and goto 2. The skip probably isn't neccessary unless you're on >2 shards.

This method works around contention by brute force and can handle a shard just disappearing. High contention will cause delays but the random time length skipping should ensure that all writes go through without too much latency. It depends a bit on how much you're willing to tolerate. It's also possible that a shard will see another shard's subsegment go back in time causing some confusion. That can be sorted by having a timestamp in each subsegment.

If you're putting a lot in the segment then checking our section is correct every tick may bit a bit expensive. If you're only using it for small messages it shouldn't be.

@mrfaul said in [Discussion] Uniformity of the world:

People would self organize, I see that as a plus.

I see it as a minus. This is a game about automation. I don't see any reason why the market shouldn't be automated. Self-organization using out-of-game methods is bad IMO.

You see there are subscription tokens which are not really traded partly because the value/afford feeling is off.

Subtokens are P2W. They let you take outside of game resources ($$$) and turn it into an in game advantage (credits). Now obviously the "W" part doesn't come without good code. I don't really see sub token trading as a major part of the game and I certainly don't think it should be promoted further.

You seem to be implying that trading is not a thing.

Trading is very much thing right now. You don't want to have to rely on getting the perfect balance of minerals especially given you have a varying need for boosts. Market is heavily used.

New resources sound exciting. My ideas for what it could be:

1. A resource that we could break down into minerals (eg. "unreact"). That would give a dense source of minerals that would require some processing to make useful. Doesn't seem that "special" when it comes to creating interesting areas to fight for.
2. Unique boosts that can't be created. A fourth tier could be added that's even more powerful but also rare/non-uniform and can't be reacted to. That'd certainly create resource wars.
3. Some sort of new form of generic resource that could be used to buff structures. For example towers could use this resource optionally alongside energy to increase damage output by 1.5x. Seems to overlap a bit too much with power creeps.
4. ???

@mrfaul said in [Discussion] Uniformity of the world:

Well most of them are only a option if you are high level anyways which is really sad.

A very small proportion of the game is spent under RCL 8, and anybody playing for multiple months is likely to have multiple RCL 8 rooms. I don't think changing low level rooms is needed: while you're leveling your rooms as a new player there are many new things to code up. I don't really see the need for low level boosting: you can either reinforce from another room, or you're unlikely to have code well developed enough for boosting anyway.

The game feels currently very linear without a lot rewarding for trying different options.

For non-linearity and "trying different options" you really have to have a combination of locking people into decisions and outcomes of decisions not being clear cut. That's hard in a programming game. We have it in the code-design, and to a lesser extend room layouts and combat strategies.

so you add a another unnecessary level of complexity to the game instead of refining the current ones

They've added a lot of refinements over the last few updates (eg. 64 rooms in pathfinder, IVM). The complexity they've added (eg. tunnels/tombstones/pull/unboost) isn't really that complex and you really don't lose much by completely ignoring it.

You can boost your creep to ridicules stats, that they are able to just crush any unboosted creep.

I do kinda agree that boosts should be less powerful, as should towers, as should repair power. I'm not sure such a rebalance is needed though and I'm also not sure what I have in mind is what you have in mind...

This would be nice.

I do think it'd be nice if shards had different colour schemes. Seems like an easy change to make things a lot more visually interesting at least.

As far as other diversity, I'd be all in favour. One of the "problems" with screeps is that there's not much to fight for. Resoures are not scarse and it's not too hard to find space. If 25% of sectors (for example) held some long lasting (say, 1 million tick) unique resource it would be interesting to have wars to fight for control of that resource. Power/space/time/energy "relics" would all suit and would allow players to actually compete for some bonus.

RIght now we have centre rooms and we have highway rooms, both providing resources that could be fought over, but IMO they're not really valuable or unique enough to be interesting.

It's actually 1.0. It's also the minimum.

https://github.com/screeps/engine/blob/f745904e856f8b6648642dbfc509a96044a5e147/src/game/rooms.js#L18

https://github.com/screeps/driver/blob/ecb3f12e20e120ea9845e00c385367e4fc704354/lib/path-finder.js#L92

@camenni said in Hostile 'tunnel' construction-sites can't be removed:

FIND_MY_CONSTRUCTION_SITES won't work on roads though

Are you sure? I've not checked, but I'd expect it to work.

@wtfrank said in Make trading subscription tokens more profitable for everyone:

You can't even pay to win. Unless you buy a top-tier code-base off someone...

I agree and I disagree. I think is you're able to develop a reasonably good codebase on your own you can get to the top of the GCL/power rankings much more easily if you're willing to sell sub tokens to buy energy/power. Similarly, a reasonably good codebase can much more aggressive with boost usage as those boosts can just be bought from the market.

Now, the market isn't huge so token prices are high. The more it gets used the lower the prices will be I guess. I personally see it as a bad thing rather than a good thing and I don't want more ways in which players can turn money into in-game advantages.