Indie Stone logo

News & Dev

TALES FROM MP

February 3, 2014

Howdy survivors. Time to talk MP shenanigans and update plans. Let’s go, go, go. Pick up planks, hammers and as much food as you can. We’ll meet you at Twiggys. If you’re not there by nightfall we’ll assume you’re dead.

Build 22 (23)’n

First of multiplayer will not be in the ‘next’ build. The next will be the aforementioned ‘Build 22’ (now, ahem, called build 23 since build 22 didn’t cut it, and us having a series of build 22’s was driving the testers crazy!) , which will be released to Steam beta as soon as humanly possible. Multiplayer, whenever is is released, will be released in a separate branch initially, and will likely have single player disabled in the short term, as its home to the still unfinished new meta spawning system. It’s almost certain that the next SP will hit before MP, however this should be a sign of MP being close not the next build being far away.

Yup build 22 is taking its time, we’ve still got issues to resolve, particularly some save compatibility issues we’re concerned about breaking older saves with. It’s in the works though and despite the push for MP a new version of build 22 (now called 23) has been in testing over the weekend, and we’re hoping the next one will be able to go out for wider testing – we’ll let you know.

 Multiplayerin’

We’ve got a slightly wider player-base for our internal tests running at the moment, the fruit of which you can see in Mathas’ recent videos. Click thru for part 2 because ads + awesome youtubers /streamers is a nice combo:

If you’d like to see what happened next at Knob Mansion and the continued zany adventures of one Flachzang, you can see more on Twiggy’s later MP playthrough here:

As well as many enjoying watching the videos in preparation for the release, a few have expressed frustration at seeing the game being played by others and not having access yet. We know it may be frustrating to not have access to the build yet, but hopefully you can see there are still issues to resolve and something as big as multiplayer we want to make sure it goes out with no server instabilities or need for the clients to log and relog every 5-10 minutes. As of right now we’ve kind of fixed the black loading bug, and the build is generally looking really stable for periods.

However there are still clearly deeper problems where connections are seemingly occasionally getting mixed up and the map streaming goes bananas, and it can have some nasty consequences for the world:

 

PZtreesaredumpsters_zps8d48dc72

We’re so close we can taste it, and after fixing the black border bug up, were hoping there was a chance we could have been releasing a public test today until this issue popped up. As soon as we have the server solid as a rock, it’s more or less time though, so trust us that it IS close. There’s no telling how long this will take to fix however, so can’t offer any solid dates. We can’t put this out there with server destroying bugs in there, but the list of things we need to fix before it’s considered ready for the public are down to the 1 or 2.

On the whole (except a disastrous day Saturday where about 20 things decided to go wrong in a new test build at once, almost all resolved now) the tests have been remarkably close to a releasable state. We have farming and fire synced up now, as well as sounds, meta events, rain and rain barrels. It’s really exciting to see it all come together so fast!

A few common questions:

How will sleeping work?

A lot of people have asked how the sleeping will work, and while the first version will likely just forgo sleeping completely, future builds will have a ‘nap’ option that will allow you to take a short sleep that lasts just a bit longer than a single player fastforwarded sleep would. The actual sleep option will be used as a way to log off safely.

How will we deal with PvP and anti-griefing

PvP, as you can see from the videos above, is having its first tests, and it works pretty great! We have some solid ideas for once the NPCs go into the game, that if a multiplayer has them turned on (and we hope they most often will be) the existence of NPCs, the initial ambiguity between human and AI, as well as having to deal with the NPCs reactions and personalities, will help curb griefing to the point where PvP could almost always be left on and unrestricted. However as we’ve stated, the default for the server will be co-op, and PvP will be something a server admin needs to activate. This is as much a statement about co-operation as it is a practical choice. As we develop the system further we have plans for a ‘safety’ button to avoid accidental friendly fire, which is all too possible in the current build, as you may see from the videos!

Loot / Zombies

Another thing that’s become extremely apparent is that both loot and zombies don’t take long to diminish in Muldraugh, at least, and we may need to extend a few areas of the game earlier than we originally planned in order to keep the servers sustainable in the short term. Note by ‘we’ this for the most part will not include Lemmy who will be busy on NPCs at this stage!) Since we’d reach the stage where the towns are picked clean and any other food has been eaten or gone bad – spawning new items / zombies in the world is not something we want to do – and since farming is currently the only source of sustainable food at that stage, we will be bumping trapping and hunting up the priority list in the coming month. This will also have serious ramifications for the types of mods that will be makeable, adding support for modders to add AI driven entities to the game. Suddenly those total conversions adding mutant / special infected / dalek types aren’t too far away.

Another potentially huge benefit to players, particularly 32bit people with a good amount of ram on their PC is the ability to host a server through the game and effectively play single player with a local dedicated server running. Due to the requirement for the entire map to stream in/out for many players, and to be leverage to use more memory to do this, it will be able to make zombie migrations work in a much more realistic way, where areas of the map are streamed in around huge hordes as they travel from the sides of the map to up the zombie populations. Seeing one of these roll through town is pretty scary, and they would obey all laws of construction everywhere on the map.

We will be looking to integrate this new map streaming system into the clients at some point in the future, the main issue being potential performance hit that a non relative streaming map would cause (server only runs at 12fps by design so can deal with it) but this is an option for those who want a more persistent world in terms of streaming, until that can happen who may see further performance improvements even over build 22 since the client is left doing a lot less work, yet reaps the smoothness of 30-60fps. That’s not to under sell the amazing improvement to FPS the build already has from EasyPickin’s optimizations a while back!

NPC phase initiatin’

We’re already prepping the ground work in the multiplayer branch to begin porting over the NPC code to work with new ‘NPCPlayer’ types, who will derive all their behaviour from the multiplayer/network communication enabled player. Once this is done we will be able to transmit NPCs behaviours easily over the internet, as well as automatically gain everything implemented for players with the next iteration of the NPCs. We can feel it all bubbling away again, and with the new meta-manager (Sadistic AI) and various other things are starting to feel the NPC train running again!

 Rezzin’

Although we won’t have a stand at Rezzed this year, we will be in attendance and organising community meet-ups on as many days as we can. If you’re coming along then post in this thread and we’ll get some cool Zomboid fun going on down Birmingham way. Can’t wait to meet you!

Related Posts