What's up this week? 11/1/12

They surely dont know it for sure, but they should have plans for what kind of hardware an average server will need to have. Like "We want the server to be some kind of more or less affordable machine" or "40 players will only be possible if you pay 500$ a month for the server". Which would be only 6.25$ per player.

I think you'll be looking at similar requirements as SupCom for that, maybe slightly less, but it probably won't be groundbreaking since there's multiple planets involved. The only advantage is that you'll be able to take the most powerful machine and let that do all the calculations, allowing you to (in general) play bigger games.

40+ games will probably still be reserved for either (top+1)end gaming pcs of that time or servers.

A "normal" game with your friends should run fine on a normal box. But remember one of the primary goals of the game is scalability. In a way it sounds like you are suggesting that we don't make it possible to run big games because that would require a powerful server. Regardless, a normal game with 4-8 players and a reasonable unit count will be fine on a regular machine. Remember when you are playing single player it all has to be hosted on your local box anyway. This is one of the reasons a dual core will probably be the minspec, so we have at least on thread for the client and one for the server.

The flow field is great but please keep/ add the ability for large units to crush small ones and not have enemy units flow around EACH OTHER. Blocking and crushing is importent, but extra path finding is better then too little

It's also a very compelling reason to emphasize "online only" play, because hosting a proper sized game may be too demanding for the average machine.

Are there any plans for distributed style hosting, ala LAN play? TotalA seemed to have a thing where each computer hosted its own AI, leading to mine always building bigger and better than my friend's puny machine. Something like this would be too fragile or simply demand too much bandwidth for online play, but it could be very powerful on the local level. It lets a bunch of average machines host something much larger than a single dedicated server might do.

a question concerning self-hosted games.
Can the server for one game session use more than one dedicated box? e.g. in a LAN scenario there might be multiple people who have some older boxes at hand, where each in itself wouldn't be strong enough as a single dedicated server, but as a group might deliver enough computing power (if used in parallel).

I am going to jump in and say no. This would require completely differet architecture that uber aren't building (to my knowledge). It is a heck of a lot of effort for what is ib the long term not really a reward at all.

the given example represented an use case for the little man.
In general terms the reward for multi node server design would be scalability.
that multi node design is different from single node design is obvious.
Since large online servers are often enough more cluster than single system servers, I just inquired if this the case here and if it can be replicated for a local setup.

The question is pretty much a matter of how much bandwidth and latency are needed between separate host nodes. It may be completely unknown at this point, which is perfectly cromulent. There's still a lot of core decisions to be made.

If the game demands extreme bandwidth and data access between host nodes, then Ethernet couldn't deal with it. If the demands are relatively low (independent worlds, limited contact between nodes, etc.), then LAN hosting may very well be possible. It's very likely going to be a challenge to see which threads can jump ship and which can't, so expect an extra 6 months in development. :lol:

I could go into the details of the multi-node setup stuff but the current design is predicated around the idea that bandwidth is more important a resource than cpu. More in a few months.

I think one of the strengths of the way they are building the game is the scalability, I dont think that any particular way of playing the game needs to be emphasized over any other. There are a wide range of computing abilities in the community and if someone has a powerful enough computer to run a game (at a minspec of a dual core id say that encompasses a lot of people) they I dont think it should be ignored in favour of "online only" (I'm not sure if you are suggesting that certain levels of game can only be played online? I dont think you are but it kind of read like that). Yes people with larger computers, or big online servers can host larger games, but again scalability seems to be one of their guiding goals. Whether someone has a beast of a machine to host a large local game or a person with a modest machine hosting a smaller game shouldn't matter, all game sizes should be allowed locally if a machine can handle it.

From what I understand they are using a server client model, so if people were having a LAN, the computer which was set to be the server would be generating the AI opponents (if anyone knows different please correct me). That is an interesting point about TA AI, I never knew that about it.

I pray that there will be alot of Offline play stuff. Cause if there isnt.... I want my money back.

Would there be a Skirmish mode for those who just want to play offline against an AI?

Where did I say anything about not being able to play offline?

I think I may be using terminology that's too technical for the crowd here. All of this client/server stuff is completely hidden from the player if you are just doing local stuff.

No keep it up neutrino its interesting stuff.

Agreed, I love the technical info.

I've been enjoying the technical updates! Even the stuff that's over my head, it's always cool to hear a pro talk about his work.

It's also gratifying to know that I've invested in a team that really knows what they are doing. Also, the massive transparency of this production is a whole new kind of business, this production is innovative on so many levels

You didn't. I was disagreeing with a comment bobucles made about emphasising online play which is what I suspect brandonpotter was responding to in kind.

Definitely keep up as you have been, I don't think it is too technical, it is really interesting : D

I really enjoy all the technical info you're giving us about PA. For all the other Kickstarter games I've recently supported, they tend to talk about art and story and graphics and other boring things like that. Building a new engine to solve the issues of a completely new way of playing an RTS is a lot more interesting to me.