ScrotieMcB vs Desync Dev Manifesto (Massive Wall Map of Text)

"

Chris wrote:

The changes to map generation so that the client doesn't have the entire map in memory will be some time in their implementation. I acknowledged the issue, and scheduled a fix, but it's large enough to not be on our horizon for the initial release which is rapidly approaching. Iin the meantime we have to rely continuing to ban people who run maphack software, rather than making their software less effective. It's a change I'd desperately love to see in, but is less important and more time consuming than other changes we have to have in by release. Such as better sync code.

I don't think you should do that, guys: creating even more client<->server traffic as a means of countering Map-hacks.

it's a delicate balance. a construction beam with more than one way to fall off, if leaning too far in either direction. obviously, more client responsibility (and less traffic) = easier to hack/cheat, but more traffic = more things that can go wrong.
you will likely need to substantially upgrade the error-handling part of your net code to counter the new situation, and even then there are some things you can't foresee.

personally, I prefer my Client does hold the entire map, and let the cheaters knock themselves out.
I don't cheat, so I really don't care.

"

CharanJaydemyr wrote:

"

johnKeys wrote:

I don't cheat, so I really don't care.

I've conveniently and subtly pointed out why your argument doesn't hold much water from the devs' point of view. :)

note that sentence started with "personally", to indicate an opinion :)

the paragraph before it however, is not an opinion but a simple fact.
I think Chris knows this very well, but it was still worth pointing out.

the middle of the "beam" is where one ideally needs to be, but if I were to choose which side to "fall" from - I'd rather give the cheaters this one opening, than risk a potentially huge re-factor of the code, and a whole lot of other - new and unexpected - "trouble".

If the servers can handle it, which I'm sure they're taking into account, then they should absolutely cut down on the ability to map hack.

"

Septile wrote:

If the servers can handle it, which I'm sure they're taking into account, then they should absolutely cut down on the ability to map hack.

think of the server as one end of a rope, and the client as another.
the rope itself is where most of the **** can and does happen, especially if it is quite long.

my personal approach - as someone who does these things almost daily - is to scale that rope as few times as possible, unless the system specification/requirements demand otherwise.

Chris, thank you for taking the time to reply, especially with your busy schedule. I'm always amazed by how the community gets the personalized touch from you; don't you have minions or something for things like this? :p

"

Chris wrote:

ScrotieMcB:

I can't respond with a point-by-point reply to your analysis of my sync article, but I believe from your reply that I must not have explained certain points well enough in my initial article. I can see that you have a good understanding of the general case of why maintaining server-authoritative sync across latency is difficult, and you have a clear passion to want Path of Exile's sync code to be improved, but your criticisms of my article show that I have not properly communicated how and why our system works as it does. I honestly feel that people are unfairly critical of this system considering it's probably one of the most advanced ones that has had to be made for an action RPG of this type. While it certainly has its flaws and can cause some terrible gameplay experiences, it does allow PoE to be one of two Action RPGs that allows zero client manipulation of combat results.

The point is somewhat moot though, because we do have substantial changes coming up to deploy to the production realm in the coming weeks. I'd rather get those done than spend ages arguing about semantics :)

And I'd rather you get those done than spend time arguing with me.

However, the one piece of feedback I'd like to give on such changes is: try to go for light, efficient, lean. The main thrust of the opening post was that PoE has a relatively bulky, inefficient resync system at its core, and to make up for infrequent global updates uses some kind of trigger set — which I'm sure is "advanced" and sophisticated and all (and I'm not being sarcastic, you'd have to cover a lot of different situations there and programming it would be a nightmare) — to compensate for this by providing targeted updates for information the client/server consider to be crucial. A relatively net-savvy player can get a feel for this just by playing the game; the way in which desync happens, and which events cause the client to resync, show evidence of an extremely trigger-reliant system with a low-to-no non-trigger information update rate. However, focusing on "lean and mean" things like entropy could allow for less bulk, more efficiency, and thus allow for more frequent updates of information — based not on triggers but just because so many milliseconds have passed — without totally killing the bandwidth. Perhaps less clever (in terms of trigger conditions), but also perhaps more efficient.

Although I guess you could also do it the hard way. Results are what's important here, and if the GGG team can make it work just building on the old architecture... whatever works.

"

Chris wrote:

The changes to map generation so that the client doesn't have the entire map in memory will be some time in their implementation. I acknowledged the issue, and scheduled a fix, but it's large enough to not be on our horizon for the initial release which is rapidly approaching. Iin the meantime we have to rely continuing to ban people who run maphack software, rather than making their software less effective. It's a change I'd desperately love to see in, but is less important and more time consuming than other changes we have to have in by release. Such as better sync code.

I definitely agree that better sync code is more important for release than fixing maphack. However, I also think that release will lead to a major upswing in hack/bot use in general, and thus maphack in particular. I can get behind a postponement past release, but if my predictions are correct I don't think it could stand to wait much past that.

"

CharanJaydemyr wrote:

"

johnKeys wrote:

I don't cheat, so I really don't care.

I've conveniently and subtly pointed out why your argument doesn't hold much water from the devs' point of view. :)

To be honest, I don't have a problem with it being used by players so much, because the main use for that would be competitive racing, and streaming makes it very hard for those in the "big time" to utilize the cheat without detection. I do, however, take issue with it being used by bots; maphack improves their pathing efficiency, and thus their bot-farming, and thus contributes to RMT, before eventual detection and ban, at which point the botter simply makes a new account.

i liked your post ScrotieMcB

one thing you might have missed is MONEY. most probably this way of handling things is cheaper on the server side and on the pipe side. 1KB saved up there and there DOES ADD UP esp when pipe is paid 'per max daily use' or similar billing method (not a flat rate). same with server power. these small 'sync' tasks etc do cost CPU power and in turn - money. money that is always a problem doing indie-stuff and reason why most indie games are single player or peer-to-peer multiplayer type of games. servers do cost a lot.

that ofc might not be true at all, but it is hard for me to find out another explanation why it has been designed that way and why it hasnt been changed since. 2 seconds desync? play melee and see that this can be as long as 4 seconds. dont believe me? check streamers esp the 'death videos'. some desync situations are hilarious.

"while it certainly has its flaws and can cause some terrible gameplay experiences, it does allow PoE to be one of two Action RPGs that allows zero client manipulation of combat results."

I can not understand how a game dev can state this and be fine with it.

Let me tell you what are the priorities for most gamers:

1. Non-terrible gameplay experience
.
.
.
.
.

.
.
.
1000. Zero manipulation of combat results.

"

Odoakar wrote:

I can not understand how a game dev can state this and be fine with it.

Because they're working on it, geez!

In the mean time we can all vent with our sarcastic comments like "I can not understand how a game dev can state this and be fine with it" while they feverishly work under the hood, pretending nothing special is going on.

Desync is pretty sucky, and I fully endorse raging out the devs for not fixing it faster.

I also fully endorse realizing that that rage will probably not change terribly much about how GGG works towards solutions, because they're probably already working at it as hard as they can in their own way.

The best thing about Scrotie's post is that he actually provides practical suggestions for solutions instead of mindless rage, which allows Chris to engage in some constructive discourse that reminds us that yes, they do have some big stuff planned, rather than giving up on desync because of what sounds like excuses from the manifesto.

Last thing: I think a lot of players would appreciate more transparency from GGG about their work on desync, but I do understand there are quite a few reasons why they wouldn't want to reveal anything (get users hyped up for a fix, only for that fix to turn out to be a big dud)

"

Odoakar wrote:

"while it certainly has its flaws and can cause some terrible gameplay experiences, it does allow PoE to be one of two Action RPGs that allows zero client manipulation of combat results."

I can not understand how a game dev can state this and be fine with it.

trade-offs happen all the time, Odoakar. it's an integral part of the process of developing something, be it a game Chris and the gang make, a piece of software my team writes, or even a road or a bridge built to connect your house and the city hall.

I respect Chris for sharing his way of thought with the community, instead of just going "shut up and pay money" or lying.
they don't make game devs like this any more :)

GGG are "green", but I do believe the end result will be more than satisfactory. at the very least, they will do all in their power to make this happen.
and if I can - despite all my QQ - be positive and trusting for once, so can you.

Yes, I agree, but this shouldn't be a trade off that one accepts. I don't know about you, but for me non-terrible gaming experiences is way, way more important than the fact some douche is using client manipulation to cheat his way though the game.