New game engine and client?

[Hal`cyon Sskyler]

Did you read what thread your posting in? this is a thread about writing a new engine. You asked me if i know anyhting about how horizons works and i don't, but it doesn't seem like you don't either so don't act like you do. Also if you haven't learned it yet don't believe everything db says.

The thread is about the client rewrite, allowing instancing would require a server code rewrite and is not at all related to the subject at hand. I'm trying to keep the thread on topic.


Instancing is a buzz word. Something put out to add to the bulleted feature list as a way to gain player interest when the players reading never fully understand how it works. For actually limiting certain players to fight in certain areas to control difficulty/rare loot drops, the game already does that to an extent. The peak of storms is a good example, only allowing so many bipeds to get though. As far as improving spawning of nameds and rare loot, there are other ways to do that outside of instancing. From what I've seen from in-game quests, especially the dragon quests, most of the benefits from instancing are possible through other, less complicated means like monster AI or having quest trigger spawns in multiple places.

Seriously, Tulga doesn't need to try and turn Horizons into Anarchy Online, Everquest, or WoW. If players really do love the ideas in those games so much, they've already switched to another MMO. Tulga keeps many of their die-hard players by distinguishing themselves from other games. For instance, the ability to solo effectively and never be forced into groups to enjoy yourself.

[Tryagain]

[img]/Web/Themes/default/images/icon-quote.gif[/img]oluviel wrote:Did you read what thread your posting in? this is a thread about writing a new engine. You asked me if i know anyhting about how horizons works and i don't, but it doesn't seem like you don't either so don't act like you do. Also if you haven't learned it yet don't believe everything db says.

Seriously, Tulga doesn't need to try and turn Horizons into Anarchy Online, Everquest, or WoW. If players really do love the ideas in those games so much, they've already switched to another MMO. Tulga keeps many of their die-hard players by distinguishing themselves from other games. For instance, the ability to solo effectively and never be forced into groups to enjoy yourself.

no definitely not but unfortunately what has been proven already is that the Horizons model as is does not work or more importantly is not successful. It certainly has never made TulgaAE one cent of profit even to this day. So then what formula will work for Horizons ? Copying other games isn't going to work but rather seeing what does work here and combining horizons features with features that do indeed work in other games. For each game is different and for each successful game out there the mold is a bit different.

So then... how to you appeal to the games die hard players ? and bring in new and returning ones at the same time ? Cause whats in game now this very instant isn't good enough to put the company into the black ink. This is what TulgaAE must find out. Thats what these "die hards" simply must get over. This isn't about copying other games its about finding the right formula for success cause as is now is not even close to successful and it certainly does not work period. All I hear is about well I don't want that in my game, well Horizons is unique blah blah blah. All fine and dandy..... what can be done to make sure "your game" is around in a year ? cause Tulga has NO MORE wild cards to pull out of its magic hats.

Now is instancing the answer ? no but it is part of the answer as a whole which is much more complicated than saying " Well stick in instancing and that will fix all the games issues " not even close that much I can agree on.

[oluviel]

Instancing is a buzz word. Something put out to add to the bulleted feature list as a way to gain player interest when the players reading never fully understand how it works. For actually limiting certain players to fight in certain areas to control difficulty/rare loot drops, the game already does that to an extent. The peak of storms is a good example, only allowing so many bipeds to get though. As far as improving spawning of nameds and rare loot, there are other ways to do that outside of instancing. From what I've seen from in-game quests, especially the dragon quests, most of the benefits from instancing are possible through other, less complicated means like monster AI or having quest trigger spawns in multiple places.

Seriously, Tulga doesn't need to try and turn Horizons into Anarchy Online, Everquest, or WoW. If players really do love the ideas in those games so much, they've already switched to another MMO. Tulga keeps many of their die-hard players by distinguishing themselves from other games. For instance, the ability to solo effectively and never be forced into groups to enjoy yourself.

Guess i have to reply with my own quote here perhaps you missed my point it's basically the same as what you stated

"Now if done correctly instancing is a great feature. Instancing to me doesn't have to be a secluded cube only your character has access to. It can also be spawning npcs by entering an area pertaining to your quest, triggering your npc via dialog, or a forced solo area via portal."

All create an instanced feel and due away with uneeded stress due to waiting for a quest npc. Im all for searching for quest npcs like your hunting them down, but say somone goes ahead and kills it ahead of you. Then you go and find out what his respawn time is, log out or wait on location. Sounds more like a boaring security job than a quest to me.

In the instance where you would be in search of this quest npc that is supposed to be hunted their could be trigger points in the geography, and when somone with the quest in their dialog enters this area a lootless version of the npc can spawn somewhere along its normal path and still provide the feel that your caracter is searching for it.
Now this is something i know ae can do because it already is in the game. For example theAegis lieutentant quest. You enter the area, dialog pops up in your chatbox,and he spawns when you do.
There are also several areas in the rop that this is utilized as well.

[Pharcellus]

Only problem is that isn't "instancing". That's called "triggered spawning".

There's nothing in HZ that prevents someone who isn't on the quest from waltzing up the moment your "triggered spawn" mob pops, and killing it before you can even get out a "buh...".

Instancing creates a pocket of the game world where such is not possible, since the mob spawns for you, only you (and potentially others who are sharing the instance) are there, only you can kill it. You can have multiple people doing the same quest in the same exact place, and they each will get the spawn, and won't even be aware of each other.

I think the point that Hal is making is that the designers of HZ don't want an instanced world. Not only are they more complicated to code on the back end, they also leave the world with even more of an "artificial" feel. It is silly to have 100 people run into a hole in the ground and, once inside, you never see the other 99 again, until you return to the shared area. Like Zodias said, sometimes it is nice to happen upon other folks, helping them out, ad-hoc group up with them, have them help you out, etc.

Instancing could be done for certain special quests, where you are tested, but it should make sense in the game world why you won't be able to have others join you or happen upon other people. However, as also said above, that is beyond the scope of the topic, as what is happening is a CLIENT rewrite. Instancing is almost completely a server-side feature. So, while instancing MIGHT solve problem X or give us new feature Y, it would be a completely separate consideration from what has been stated so far.

[oluviel]

i've said my peace about instancing already. Lets move onto more important things. Dungeons, capes, andpets am i under the wrong impression that these are server side issues also?

[Pharcellus]

Dungeons would probably be mostly server-side, as they require a world model.

Capes are already in server-side, they just need art assets and client-side updates to see.

Pets would also be mostly server-side. Only thing that I can think of for pets on the client-side would be their UI.

[Drevar]

Pets and dungeons are probably 50/50 server/client. Capes would be the only thing that would fully be doable client side.

The cloth engine already exists. Anchoring the cape to the character is the issue. I have yet to see a 3d MMORPG use free flowing capes that weren't somehow glued to the characters. Usually they are akwardly stuck to the feet so you never actually get that "Spawn" cape effect going by any means. Most people don't have enough CPU oomph to use that kind of effect with a dozen plus characters on screen (especially with the cloddy renderer we have now). The new PPU (Physics processing unit) from AGEIA could make this type of effect common in any game.

With a total rewrite of the client in the works, it would be really cool to know that it was being done withPhysXchip support.

A small edit since Malt responded just before me...pets would also need some updating of the client side prediction code. We already see how clunky it is any time you are following someone over a bridge and they suddenly blast off into space, and a few seconds later come crashing down through the bridge.

Drev

[Tryagain]

saw my first cape in COH last night. They have that down pat about as realistic as I have seen in a game.

Dungeons can only work in a NON instance way if the game has a world model. Im sorry but I never ever see that happening in this game unless Tulga finds an investor that will invest $10 Million plus into the game. Yes it is that much of a project and really something that needed to be done concurrently with world design back in development.

Pets as well cannot happen till the world model and some kind of collision detection is in. Or you would have pets running all over the place just trying to follow you ;) .Pets is a much more complicated issue than most people can appreciate.

As I said you want these changes done in our lifetime one of you needs to donate their lottery winnings to TulgaAE ;) . Or in some kind of timely manner at least...

[Pharcellus]

Pets don't need much on the client side, just some way of controlling their AI. They are nothing more than a mob with a player-controlled AI and their friend/foe flag set to "friend". Thus, they don't need very much on the client side to work. They need to add at least a basic world model so that they are aware of objects (rock outcrops, bridges, buildings, caves, etc) so your pet can go everywhere you can, but it could be implemented without the model; they would just walk through/below objects just like mobs do now. Yes, it would be minimal, and it would be annoying at times (you have to tell your pet to "stay" outside when you went into a building or cave or it would do so automatically), but it could be done.

Instanced dungeons are one of the primary uses of instancing. Instancing has nothing to do with a world model. Not sure where you are coming from there, Grim.

[oluviel]

damb lol so what will a new client actually do?

[Scarn_de_Diabolic]

Hopefully allow for 30 or more people to be in the same area without 2 fps.[Y]

[Tryagain]

Pets don't need much on the client side, just some way of controlling their AI. They are nothing more than a mob with a player-controlled AI and their friend/foe flag set to "friend". Thus, they don't need very much on the client side to work. They need to add at least a basic world model so that they are aware of objects (rock outcrops, bridges, buildings, caves, etc) so your pet can go everywhere you can, but it could be implemented without the model; they would just walk through/below objects just like mobs do now. Yes, it would be minimal, and it would be annoying at times (you have to tell your pet to "stay" outside when you went into a building or cave or it would do so automatically), but it could be done.

Instanced dungeons are one of the primary uses of instancing. Instancing has nothing to do with a world model. Not sure where you are coming from there, Grim.

umm no for pets to work period there needs to be a concrete way of telling the pet where the ground is and its relation to the landmarks around it. In order to do this you WILL need a world model otherwise I promise you Tulga would have put them in already. They aren't dumb thats for sure ;). I can see many ways of exploiting the use of pets if they were coded like the mobs.

But what I was saying is that instancing can be done WITHOUT a world model. Its the one thing they could add that wouldn't require the tons of work a world model would need. But they would need better server engine thats for sure.

But back to original topic you know at this point... I would be happy if the game even came close to performing like COH and WoW and the older MMO's. Its having to deal with the degradation of performance that EVERYONE has to put up with that sometimes make the game more of a chore to play than fun. Content I could be content waiting for if horizons wasn't such a poor performance issue.

[Pharcellus]

A new client would (theoretically) dramatically improve performance.

..and, no, Grimlik. Pets can work without a world model just fine with the restrictions I suggested. Every creature in the world at least knows where the ground is. The artifact that you are seeing where other toons, monsters, etc "fall through" a bridge, etc is because those are objects, NOT terrain, and the server has no knowledge of them (because it doesn't have a world model that models them, go figure).

From observation, the existing "world model" knows only two things, X/Y position and altitude (not the same thing as "Z", though). To mobs and NPCs, the world is completely 2D and flat. Aggro is determined by simple 2D distance calculations, with an altitude check. Mobs see players and move in 2D towards/away from them, even up sheer cliffs (which they don't know exist; remember, to them, the world is flat). As far as the terrain engine goes, this isn't a problem, as the client scales the 2D world to the heightmap of the terrain. However, when you introduce an object into the mix that only the client knows about, like a bridge, the server still operates in 2D, but the client can't automatically tell whether the mob is supposed to be on the bridge or under it, so defaults to the terrain mapping, and thus mobs fall through it.

A similar problem exists for PCs running across the bridge. The prediction code is wonky and tries to figure out where to put everyone else. The client KNOWS where your toon is all the time, and handles your collision with all surfaces, be they object or terrain, just fine. The problem is that it is not told, exactly, where your buddies are, so it tries to guess and often does so... badly. They've worked on that problem some in the past, so you don't see the falling through bridges and floating up through cave ceilings problem much anymore (at least it isn't as noticeable, but their workaround is far from perfect).

However, for the mobs, the problem is that we have added another path. The players can walk onto the object, or under/into it, but the mobs don't know it is there, so they don't know what to do. I think, in most cases, they think you are flying, as your altitude climbs even when they think they are standing right on top of you.

Pets could be implemented now, it is just that they would suffer the same problems as mobs do. They couldn't go into caves, upstairs in buildings, towers, etc, or onto rock outcrops. They could be coded with "forbidden regions" where they will simply stop following you and go into an "idle mode" when you enter that region, waiting until you come out of it. The big problem they would have to work out is crossing bridges, which is likely one big reason why they are waiting for the implementation of a minimum world model that is aware of the objects, for both mobs and pets.

[Milandra]

Bump
Any news on this ? [:$]

[Varangaard]

Hmm, they said one reason to segment the board and move posts out of general was this section auto-magically disintegrating all stuff older than three months. Verdict: "they fail it".

I haven't hung on IRC much, but I don't think there are any news (I assume you mean client, as it is the main topic of the thread).

[imported_Frith-Rae]

..and there has been "upgrades" to the current client...

..wether or not those increased stability for any one person individually, is up for debate - it did for some, it didn't for others, it made it worse for even others....

..so the only way to iknow wether any one former player would be helped would be to login and try it ;)

(I know it helped me, I can now use 3 more towns I couldn't go too since the merge because of lag/ccrash immediately upon arrival..HUAZZAH!)

[Goriax]

I asked david a little while ago whether they were still working on the new client, and he said: (paraphrased) Yup, we split the client team into two teams. 1 working on the making the current client more useable until the new client is done, and one working on the new client. This work has been going on for about a year now.

That's good news. Really good news. When investigating rewriting the main client at work, one thing which became apparent very soon was that it was going to be hard for version 1 of the new client to have every single feature of the old client. I wonder if that will be an issue here too. Of course our client probably has more features then the HZ client including some which hardly anyone ever uses so maybe it isn't the same. I wonder how bad the flaming would be if the initial version of the new client only supported 90% of the features of the old although perhaps the old would remain available for awhile and perhapsthe newwould support all the heavily used features.

But that's all wild speculation on my part. It's good to hear they're so far along. Now I"m curious what engine they're using.

[bright]

for me at least, the client wouldnt be bad now. i get on average about 50 in the country and i hit a record of 30 in taz, but man port a few times and im lucky to break 20, and the lag spikes are simply horrible. go to enter the cave for the lairshape trainer 40 40 40 get to the bottom SPIKE 6 6 6 (2 minutes later) 30 30 30 honestly though if client stability was like most of the games out here this game would quickly be revived. still my all time favorite. has the coolest things the nicest people and CAN be the funnest<especially considering theres some decent RP>but performance still is nowhere up to par with where it should be this far after release.

Though the event have started to roll out again so that does mean that things are rolling. So, keep up the good work. hopefully this will solve all issues with poor performance and finally give HZ the respect that it deserves.