Haven 2.0: "Hafen"

[Oddity]

@oddity Thank you for the updates notes now and then.

No problem. Still thinking about pursuing this idea:

Someone should put together a list of features that have been mentioned by jorbtar that they intend to include in Hafen.

Oddity, you up to the challenge?

I've actually been meaning to do something like that. I want to make a searchable list of things the devs have said, with tags so you can search for their thoughts on a topic, eg. Hafen, combat system, etc. Haven't done anything yet, though. Best thing you can do is just skim over my Hafen thread.

[Oddity]

Holy crap, look at this video: viewtopic.php?f=6&t=37974&start=1790#p512587

I'll just leave this here.

[LadyV]

Here oddity Ill help.

Hope there's more click reduction measures along the same lines.

This feels like luxury compared to the old system.

I know; I'm very worried about a whole new generation of Haven players growing up not knowing how hard life can be. I'll probably just have to remove it again.

This might be slivovitz talking, but I love you Loftar.

Sooner or later. ^^

[Oddity]

Probably not surprising to anyone who pays attention, but I thought it might be worth posting anyway:

(09:54:19) <biohazard504> when is kickstarter for hnh 2? someone mentioned it earlier?
(13:15:04) <loftar> biohazard504, it is extremely unlikely that we'll hold a kickstarter.

(15:04:35) <loftar> Sorry, I don't consider myself having time for that right now.
(15:04:49) <deus_> ill just leave it here then
(15:05:04) <deus_> yeah gotta make room for dat animu
(15:05:49) <loftar> I was trying to investigate I/O latency, actually.
(15:06:01) <deus_> theres no documents on this?
(15:06:42) <loftar> Surely, but I haven't found anything that is applicable to my context.
(15:07:01) <deus_> what are you trying to figure out?
(15:07:09) <deus_> how much keyboard lag in your dota game?
(15:07:43) <loftar> ^^
(15:08:06) <loftar> Well, the vast majority of lag spikes in both Haven, Salem and Hafen is related to I/O latency.
(15:08:13) <Oddity> Interesting
(15:08:18) <deus_> I/O where
(15:08:22) <loftar> Where some simple filesystem operation, like rename(), takes a long time.
(15:08:54) <deus_> id assume its seek time
(15:09:08) <loftar> No, those times are too short to be a problem.
(15:09:14) <loftar> Rather, it's about queue congestion.
(15:09:47) <deus_> did you assign a latency value to each different op type
(15:10:27) <loftar> No, well, the more basic issue is that I think the vast majority of them shouldn't have to block on disk I/O at all, since there's so much RAM free that I think the operation should be entirely bufferable.
(15:11:17) <deus_> are you trying to figure out which ops cause latency or why they do
(15:11:24) <loftar> Well, I'm fairly aware of that.
(15:11:30) <loftar> What I'm trying to do is find out why they take long.
(15:11:39) <deus_> enumerate ops
(15:11:47) <loftar> Mostly rename().
(15:11:54) <loftar> Sometimes creat().
(15:12:19) <deus_> any branches?
(15:12:32) <loftar> What kind of branches are you talking about?
(15:13:08) <deus_> anything that would cause branch prediction if it is indeed buffered
(15:13:29) <loftar> What does branch prediction have to do with I/O buffering...?
(15:13:51) <deus_> i cant see what you see...
(15:14:02) <loftar> Well, the problem is not CPU-related.
(15:14:37) <loftar> It's mostly due to some of those operations having to carry out physical I/O while the disk is busy.
(15:14:47) <loftar> And my point being that I don't think they should have to do physical I/O.
(15:14:57) <deus_> did nsa break your firmware?
(15:15:05) <loftar> I think they should be able to deal with their business in the buffer cache.

[Oddity]

viewtopic.php?f=6&t=37974&start=2040#p514184

So far HnH has proven that it is kinda of impossible to impose such "soft" standards on people in an online game, as they do not really need to fear any consequences if they do not follow a certain kind of code.

To be fair, I think there is a case to be made for the sometimes-made argument that other players in one's area is too much of a liability and too little of a benefit to keep around. If there were more real benefits to having Farmville-players around oneself and they presented less of a liability, I think it would be more reasonable to expect at least some of the more combat-oriented people in the area to protect them.

viewtopic.php?f=6&t=37974&start=2040#p514189

regarding climbing cliffs, how does the topography of the new version work? is it voxels or something?

No, the map is still 2.5D with one layer where each "vertex" has height information, and cliffs "arise" due to the height difference between two tiles being over a certain threshold.

can there be "stronghold crusader" type walls that people walk on with ladders and stairs?

I ambition to enable that sooner or later (depending on other priorities), but it's not going to be possible initially.

Have you considered any concrete measures for Haven 2.0

There aren't much in terms of concrete mechanics yet (though villages will likely work better than previously), but the terms I tend to think in revolve around mechanics that try to correspond to "active playing" (the term being intentionally loosely defined) of a character being rewarded in various ways. I imagine that if even nub characters, through such rewards, are able to bring to the world valuable items that can bring benefits to more advanced characters (or advanced characters bringing rewards to other, differently profiled, advanced characters), that will bring more reward to keeping them around.

[Granger]

Probably not surprising to anyone who pays attention, but I thought it might be worth posting anyway:

(09:54:19) <biohazard504> when is kickstarter for hnh 2? someone mentioned it earlier?
(15:05:49) <loftar> I was trying to investigate I/O latency, actually.
...
(15:08:22) <loftar> Where some simple filesystem operation, like rename(), takes a long time.
(15:08:54) <deus_> id assume its seek time
(15:09:08) <loftar> No, those times are too short to be a problem.
(15:09:14) <loftar> Rather, it's about queue congestion.
(15:09:47) <deus_> did you assign a latency value to each different op type
(15:10:27) <loftar> No, well, the more basic issue is that I think the vast majority of them shouldn't have to block on disk I/O at all, since there's so much RAM free that I think the operation should be entirely bufferable.
...
(15:14:37) <loftar> It's mostly due to some of those operations having to carry out physical I/O while the disk is busy.
(15:14:47) <loftar> And my point being that I don't think they should have to do physical I/O.

I guess fsync combined with data written but not yet on disk could force all dirty buffers being commited, leading to blocking on the rename() or create().

[Oddity]

by jorb » Sat May 16, 2015 3:50 pm
Oh, no. Not at all. No bumps. Very smooth sailing.
We'll have something for you before long. We are working toward announcing a release date.

[Oddity]

FUCK YEAH A VIDEO AND ETA: LATE AUGUST 2015

viewtopic.php?f=6&t=38988

[Oddity]

We're still debating the ifs, wheres and hows of building ovens inside houses.

viewtopic.php?f=6&t=38988&start=140#p517277

You seem to completely misunderstand what I'm talking about in the post you quoted. I'm not talking about the game itself, obviously, since that's what we're currently working on and trying to improve. As we've stated on multiple occasions, we consider the game we're releasing in August to be Haven & Hearth -- the natural continuation of the game currently running. It is going through the same kinds of changes that it would have done if we had simply continued working on it without doing the excursion with Salem. If anything, the experience with Salem lets us avoid some things that we would perhaps have done otherwise that turned out to be less-than-optimal.

What I'm talking about in the post you quoted is, indeed, the current version of the code. The code itself is outdated, obsolete and replaced with better, leaner and more capable constructs. It is purely a remark on the technical aspects of the implementation, not the game itself.

[Oddity]

<loftar> I did recently parallelize the Hafen client to be able to use at least two cores.

(11:42:44) <loftar> Hmm, well, to begin with, the mapgen is still pretty much up in the air. The testing world right now just used a very slightly modified Salem mapgen.
(11:43:07) <loftar> But, well, I generally do think that rivers are *too* pervasive in Haven right now.
-
(11:44:47) <loftar> Well, I think it is somewhat fair to say that I would want longer but fewer rivers.
(11:45:05) <loftar> And with more reasonable systems of tributaries.
(11:45:22) <loftar> Rather than just a mesh laid onto the map.
-
(11:47:54) <ArvinJA> loftar: Any thoughts about sieging so far? I don
(11:48:11) <ArvinJA> I don't need details, I only want to know if that's been a hot issue
(11:48:27) <ArvinJA> specifically solutions to the old "timing" problem
(11:49:25) <loftar> Not for now. We're sufficiently busy right now that we're using the current system "by default", so to speak. We're very open to changes, however, and are often thinking about it.
-
(11:51:35) <loftar> One of the greatest... controversies we discuss amongst ourselves is to what extent we wish to "formally model" social/political mechanics.
(11:52:11) <loftar> I mean, one thing we have discussed is having a system somewhat similar to Salem, where things on a town-claim are completely indestructible until formal war has been declared on the village.
(11:52:27) <loftar> And, well, neither I nor Jorb are sure what to really think of such a system.
(11:52:45) <loftar> Part of me very much likes the very free-form and non-forced systems in the current Haven.
(11:53:04) <loftar> While another part of me wouldn't necessarily hate such "symbols" and "social modeling".
(11:53:18) <ArvinJA> loftar: Yeah, what's the benefit of that? I don't think that was how it worked socially/politically historically
(11:53:37) <loftar> Of course not, but historically people wouldn't be logged off from the world, either.
(11:53:47) <loftar> The point would of course be that it would take, say, 24 hours to declare such war.
-
(12:03:12) <loftar> I mean, one system I've long wanted to fit it somehow, is the ability to sneak past one's walls for some limited purpose.
(12:03:33) <loftar> Like to do some very minor crime, or to punish a criminal one is tracking, or something.
(12:03:57) <loftar> Just to decrease the necessity of completely razing a village just for minor purposes.
(12:05:02) <loftar> Given that, it would also be much more okay to make it a lot harder to actually raize a village.
(12:05:06) <Krikke> Ofcouse not murder as punishment, that would just break the game, but something like steal LP or something, just a thought. Something minor to punish them in some way.
(12:05:08) <loftar> The problem is how to make such a system good.
(12:05:24) <loftar> I generally dislike the current "boolean" aspect of raiding.
(12:05:31) <loftar> It being very much all-or-nothing, that is.
(12:06:11) <loftar> And I would also like it if there were greater risks involved in raiding overall.

(17:46:14) <Alg> Pls tell me
(17:46:18) <Alg> there is no more curiosity system
(17:46:30) <loftar> There is indeed a curiosity system.
(17:46:32) <Rawrz> troll (childhoodobesity) and the megam8's are one group that does do that though
(17:46:35) <loftar> Why would you not want there to be?
(17:46:43) <Alg> Because it encourages the building of Alts
(17:46:51) <Alg> You just log on, put curios on the character
(17:46:52) <Rawrz> we killed jordan coles and ladygoo while they raided Alg's place
(17:46:53) <loftar> Well, it has been modified.
(17:46:54) <Alg> Log off, repeat
(17:47:01) <loftar> So as to prevent that.
(17:47:04) <Alg> How so?
(17:47:12) <loftar> That you'll have to wait for and see.