Fixing a broken/boring game by going free to play will probably not work.

A shitty game is still a shitty game, whenver it's free or not.

We had our alpha version out for some time, while doing a range of small bug fixes and minor features, we are now starting to discuss major game play mechanics to implement, join our conversion at Get Satisfaction or our forums, we need your ideas and your feedback:
http://getsatisfaction.com/abydosonline/ideas/recent
http://www.abydosonline.com/forum/viewforum.php?f=13

I am investigating different payment models for MMO's.

I am asking this because of the problem with gold sellers, if you do not have a cash shop, there is a high risk your game will suffer from gold sellers, which means accounts getting hacked, in game spamming etc.

If you bring in a cash shop, you are favoring people with more money and thus creating an unfair advantage for people who happen to be richer in real life. On the other side, without a cash shop, those people that buys gold will probably do it through gold sellers, taking advantage of their real life money anyways.

The poll is simple, is it a viable solution to bring in a cash shop to get rid of gold sellers, with the fact of that people buys virtual currency from gold sellers anyways.

If you have any other suggestions that gets rid of both cash shop and gold sellers, it would be very nice feedback.

Subscription: Monthly fee.

Cash shop: Buy virtual currency for money (no items).

I would say tetris with pygame (2D game library for python) would be the best newbie option to get hands on on game programming.

You are completetly missunderstanding your own situation.

The problem is not about levels it is about game play.

If the game play would be fun, you would have no problems leveling 80 or 999 levels. As long as the game play is fun, you enjoy your time and continue to play.

Originally posted by ActionMMORPG

One type of server lag is related to how fast the server can tick() in terms of ticks/second.  FPS software used to tick anywhere from 32 to 75 ticks per second.  In best case, each tick it would move, handle combat and AI for every object in the game world.  That's not hard in FPS because there's not much scripting and few objects.

 

To my knowledge MMOs don't tick nearly this fast, maybe as fast as 8 ticks per second on things which are active, and as slow as 1 tick every 5 seconds for things that aren't active (zombies in the world like mobs with no aggro).

 

When a server loads up, sometimes what happens is the CPU (or database etc) can't keep up with the 8 ticks per second needed to keep things fluid...  In slowing down (fewer ticks per second) the communications to client, the client itself starts feeling sluggish and often the movement of things starts getting rubber-bandy.

 

Scalability is the method of devoting more resources to a process so it can handle the load.  Often this means spreading out over more computers so they can handle the CPU.

 

In a perfect case, client ticks at 100% as fast as it needs to, server ticks at 100% as fast as it needs to, and network can handle the updates between the two in near real-time (50ms delay or less).  if any of these doesn't happen, the player starts seeing the effects of lag (latency) or lag (missed updates).  If scalability is in place, it should detect the overload and fix it.

 

Doesn't always work that way.  I've seen video of one recently where swing to hit an opponent, and the actual hit was a good 5 seconds late.  That's seriously delayed, and no where near realtime.

Very good post.

But I also want to say that not all MMO's architecture are design in this way, for example, in my own server (www.next-gen.cc) there are no such things as ticks, or a main loop.

The server has hundred thousands or millions of small processes, which in most of the time take no CPU because they are wating for messages.

Since I am programming in Erlang, I get this implementation for free, there are similar ideas implemented in Stackless Python (Stackless Pythons stacklets are faster than Erlang processes but less elegant and as powerful powerful as Erlang processes).

If you are a DIY type of guy, you can always implent your own processess/stackless/green threads whatever you call it, by using something simple as the switch statement .

Originally posted by Coder

 

---

Originally posted by ActionMMORPG
Scalability is the method of devoting more resources to a process so it can handle the load.  Often this means spreading out over more computers so they can handle the CPU.
 
In a perfect case, client ticks at 100% as fast as it needs to, server ticks at 100% as fast as it needs to, and network can handle the updates between the two in near real-time (50ms delay or less).  if any of these doesn't happen, the player starts seeing the effects of lag (latency) or lag (missed updates).  If scalability is in place, it should detect the overload and fix it.

---

 

Devoting more resources to a process means you have to have more HW resources to devote and that the process is threaded in some way so it can be distributed. This also leads to more overhead which also consumes HW resources. So this is not always feasable and at one point you get more overhead then the actual job consumes.

"...If scalability is in place, it should detect the overload and fix it." How do you fix network lag with this? How you do you fix client side lag with this if the CPU is working at close to 100%? How do you fix server side lag if the CPU('s) is close to 100%?

"...more computers so they can handle the CPU." huh?

You are completely wrong in your statement that code needs to be threaded to scale. 

For example if I have 4 cores in my machine, I need threads to use all four of them (in this situation you are right).

If I have a cluster of lets say 16 machines with one core each, I do not need any threading at all (in this situation you are wrong).

Most likely, scaling without threads is very possible in a network of clustered servers (horizontal scaling), but the threading is only used for scaling locally to more CPU/cores on the same computer (vertical scaling).

Originally posted by xKingdomx

A lot of people lately have complain about "same old same old" with feature repeating itself in multiple mmo. I just want to compile a list of features that people want to see featured in future MMORPG, whether it be from single player RPG, other MMORPG, action games, etc

1. Rhythm based combat, where timing is the most important. This prevents the game from becoming a twitch-fest based game, while still requiring some sort of mobility and coordination from players. It will be like going out and throw a frisbee with friends, not intense dodgeball training. Notable Example: Assassin's Creed

 

2. Movement ability, this feature is so restricted in superhero genre that it isn't really being explored in other genre of MMORPG, I myself hates flying ingame, it destroys the whole purpose of environment/zone design when you just fly over stuff (and no, no fly zone is stupid, like in AION), surely fantasy or sci-fi game can incorporate movement abilities into the story without making it sound stupid, eg: Jedi can force jump, force running (sci-fi), Assassin's can shadowstep and run faster than normal (this is actually in most class description, yet to be realised) Mounts are simply made for travelling purposes, sprinting and scaling walls can attribute into combat tactics as well. Notable Example: DCUO, Champion's Online, to a lesser extent, Assassin's Creeds scaling walls

 

3. Physic's based world, with next-gen game engine like UnReal3, CryEngine 3, I cant see physics based game be hard to make. Interactive objects, projectiles, collision detections are all very simple to do with today's technology isn't it? We shouldn't need to rely on percentages to determine dodging or hit change. This year is 2011, not 2004 when WoW came out, lets show our advancement. Notable Example: TERA, GW2, DCUO, Elder Scrolls, and to a lesser extend, shooter games

 

4. Multiple combat system, each class shouldn't be only distinguishable by the class description text, your skill in role-playing, graphical animation. Classes are different, their skills are different, their roles are different, so when you think about it, should they play differently? Bow and sword/dagger wielders demands much more physical contact in their combat, so their combat will be more collision based like the physics based point noted before. Whereas Mages or magic users are much more static and concentrates on the effect of their actions, that demands a closer approach to the traditional hotkey combat. Notable Examples: Never seen anything like that!

 

5. Weapon based combat, instead of mastering one weapon or attunement, having 100 skill that we will eventually grind down to using only 10 of them, we can use multiple weapons, which each weapon classified into a specific purpose, with 3-5 skill, and it can act as a toolbox to react to the shifting combat dynamics. This idea was spawn from shooter games , how you use different guns for differing enemies at different range. Notable Example: Vaguely attempted at GW2 (I think) Assassin's Creed, to a lesser extend, shooter games

Originally posted by wizyy

Originally posted by ironhelix

I won't go near this game because the characters all look like 10 year old girls, regardless of the actual gender of the character. I frankly cannot stand to look at it, because it's disgustingly vomit-inducing. Let's not forget that it's been done TO DEATH anyway, so even if I WAS into sexualized children, I still wouldn't like it. The asian video-game art design talent pool is totally devoid of any imagination, or innovation.

You confused ArcheAge with TERA, my friend.

http://www.youtube.com/watch?v=eQwnb3TzbEg

Tera Elin race really looks to be made for pedos. 

I can understand most people does not like this kind of art (including myself, I do not like the idea of 10 years old girls acting as fighters in a serious game).

However, I beleive the creators of the game created this race for their female player base, rather then pedos. 

Originally posted by k11keeper

To be completely honest, I wish every game was like EVE in that way. Having everyone able to be on one gigantic server is freaking amazing. I don't see it being very possible in fantasy style games though. At least in the near future. Unless it's setup like GW but then it's not a seemless world.

Why do you not see that is possible, and what do you base that on?

There is no real difference between a space and fantasy game in terms of technology, other than that space game tends to spread players more in x y z while fantasy would spread mostly x and z while y (up direction) would be similar values.

But as I said, that difference does not really affect the technology to run single world MMO's.

ArchAge has an claimed to have an interesting number of gameplay mechanics.

What I do not like about it, is that it will probably shard the popluation across different servers, I would rather the one single world (like EVE).

Personally I am not fond of asian girl/managa style graphics.

These are the only down sides from my point of view.

• You do not care about immersion, meaning that you are not intrested in the world or feel a part of it (Later on, you actually admitted you felt a level of immersion but I did not have that kind of information at the time)

Originally posted by khanstruct

Originally posted by Vagelisp

Irrlicht is not an engine. http://irrlicht.sourceforge.net/faq.html

 

Unity is far more complete with nvidia physics sound, terrain editor, etc but it's not free at all. You need at least 1500 dollars in order to get "serious" http://unity3d.com/unity/licenses.

Unity is absolutely free. Yes, you can do some cool post-process stuff and asset bundling with the pro version, but you can easily make a fully functioning game with the free version. I've already done it, and I'm working on doing it again.

http://www.kongregate.com/games/Democre/starfall-frontier

Kongregate is full of Unity games developed using the free version.

The free version of Unity does not support shadows, so if you need shadows you must buy the pro version.

Originally posted by Meowhead

Originally posted by abydosonline

You are completetely missing the point here, comparing chess to an MMO with the only common factor that both are games is a moot point.

The reason for this is the purpose of the two different games.

• The purpose of Chess it to beat your opponent in a logical puzzle.
• The purpose of an MMO is to immerse the player into a virtual world together with thousands of other people. The point Anubisan was making was that the immernsion suffers if there are no consequences or risks in the actions the player does.

He didn't say 'MMOs need harsh consequences', he said that games do.  He in fact said that's the entire POINT of games.

I think you're the one completely missing the point.

Also, you made a very, very big erroneous assumption in your post.  'The purpose of an MMO is to immerse the player'...  No.  That's the point FOR YOU.  For me, and many other people, the point to an MMO is that you're playing a game.  With other people.  For fun.

You might have immersion as the #1 factor that is vital to an MMO, and while you aren't alone, that isn't true for everybody... or apparently, even the majority of people, if you look at what kind of games sell.

So I got the point, and made my own very well, thank you.  ... but thanks for trying, I appreciate the effort on your part, even if you missed two points (He specifically said that it's the point to games, not to MMOs, and not everybody has immersion as the most important... or even an important... factor in MMOs.) that are rather important.

 

Originally posted by bunnyhopper

Nicely put.

... but don't feel bad!  At least you're not the only person who missed both points. :)

Let us first define the word immersion which I used, to me it means these things:

• Feeling a part of
• Being interested in

Originally posted by gobla

Why are you writing the server in Erlang and the client in C++?

A MMO server is about the most performance intensive piece of game software there is ( it has to run the game logic for thousands of players as well as manage all those thousands of players their network connections. ) while a MMO client is about the least performance intensive ( it doesn't have to run much game logic at all. )

Yet you're choosing to use a garbage-collecting run-time language for your server while you're choosing to use a compiled language for your client. If you're capable and able to write C++ then why not use it for the server? It would lead to increased performance.

I know BSGO ( BattleStar Galactica Online ) is also using Erlang for it's server but their reasoning is that because it's a run-time language that supports independant processes they can at any time add more machines to their server cluster to increase server performance.

However as an independent developer I think you'll have a hard time constantly adding/removing machines from your server clusters from a pure financial point.

Wouldn't it be much smarter to write the server in C++ as well and take advantage of the increased performance that provides thus saving costs on the amount of servers needed and even more costs on dynamically adding/removing even more machines from your server-cluster?

I mean I'd understand using Erlang if you were aiming for fast and easy development while still retaining decent performance due to the functionality provided by the independant processes. But if that was the case then I'd expect you'd be writing your Client in for example Java.

So why exactly are you writing your client in C++ and your server in Erlang? Why not the server also in C++ or the client in say Java?

Doing horizontal scaling in Erlang is feasable for a single developer, my server already have the ability to add more servers to the cluster to share the workload of an area.  Buying VPS servers on Amazon or any other provider is not expensive these days, once you got about 1000+ paying customers I would not see any trouble buying a few servers, as a matter of fact, I already have two servers up and running Abysos Online (costs about 40 EUR per month).

As long as you use Erlang for things that Erlang was meant for, meaning concurrency and distribution, your software will be on par with an implementation using C/C++. 

However, when you do things like sequential programmering and number crunching, Erlang is roughly ten times slower, to work around this you write some parts of your programing in C/C++ and interface it with your Erlang software. Right tool for the right job is a reoccuring mantra :)

At work for example, we have the MMO server running Erlang, but the quad tree, calculating what quads that are hit depending on a position and an area of responsibility is written in C.

My preferred language for the client would have been Python, but the python bindings for the rendering we are using (Ogre3D) was a mess, the easiest solution was to use C++.

Originally posted by Meowhead

Originally posted by Anubisan

I think its despicable that games have come to the point where people die on purpose just to fast travel. Doesn't that defeat the entire purpose of a GAME!? I mean what is the point of anything if nothing has any consequences? If there is one thing I could change about the formula modern MMOs have adopted, it is most definitely the death penalty.

... man, I'm not playing chess with YOU.

'Sorry, you lost, I'm going to have to kick you in the face'

'What?  I gave up because I saw I was at a significant disadvantage, and wanted to try another game, because I think I could do better restarting!'

'... you're totally missing the purpose to a GAME.  Now duck your head, I can't kick very high.'

Most games, the penalties for losing are trivial.  In games with serious penalties, they often far, far outstrip anything any MMO has ever had for a death penalty.  Ever.  MMORPG death penalties are pretty pointless in general.  They're punishments of time, that's it.

You are completetely missing the point here, comparing chess to an MMO with the only common factor that both are games is a moot point.

The reason for this is the purpose of the two different games.

• The purpose of Chess it to beat your opponent in a logical puzzle.
• The purpose of an MMO is to immerse the player into a virtual world together with thousands of other people. The point Anubisan was making was that the immernsion suffers if there are no consequences or risks in the actions the player does.

In Fable, you could buy a title, or somehow unlock it through achivements. This was completetly decoupled from the skills you choose to invest in.

In an MMO, it could be used to give the freedom to players to choose how they are percieved by others.

Updated the original post with some screen shots of the client.

Originally posted by khanstruct

I'm not doubting your ability TO do it. I've checked out some of your stuff, and it's obvious that you guys know what you're doing when it comes to code. My concern is whether you SHOULD do it. I know fully (and permanently) interactive environments are possible, and your team may even be capable of doing it well. From a design standpoint though, this is a terrible idea, and all of your years of hard work will be destroyed in a matter of minutes.

I agree with you, there are a few things about dynamic... or should I say destructive environments that can go awefully wrong in terms of gameplay.

We have not yet discussed this in detail yet, but will probably want some sort of land ownerships, making a player have to buy or claim land before raising/lower the terrain. An insurance system for houses would also probably be a good idea.

I am not saying we are doing right, I am just saying we are trying do it, whenever it is a success or fail, the future will let us know :)

Originally posted by khanstruct

Not to be the rain cloud here, but there are some serious issues with your design. And actually, it bothers me how common and simplistic the issues are.

Fully interactive environments? This is the dream of every 15 year old kid (who has never studied a thing about game design). Your game would be live for about 20 minutes before every forest was cut down and every mountain was flattened. Sorry, but it just seems like this game is more of an unrealistic wishlist than a game design.

Don't misunderstand. I think its great that you have a solid team, and you're ambitious enough to take on an MMO. That's also part of the reason for this response. I'd hate to see a team of talent waste years on a faulty design.

Yes I understand your concerns. I can explain more in detailed about this interactive environments:

Interactive enviroments is not same as game objects with full, realistic physics.

Each game object will on creation generate an unique id based on the server the object is created on and the current time in microseconds. As long as just one id generator per server in the cluster is running this should ensure truly unique ids.

Game objects will be divided into three states:

• State 1: Unloaded, will only take space on disk (in the database).
• State 2: Partially loaded, I suspect that most of queries for an object will ask for the object 3D model and position / orientation, just these will be loaded into a memory resident database. Will take up small amount or RAM.
• State 3: Fully loaded game objects, in this states the game objects have their state loaded into the memory of a Erlang process (do not confuse Erlang process with threads) and is ready to receive state changes (this means objects can be interacted with). Will take up a larger amount or RAM.

As an example, imagine a city with lets say 500 buildins spread out, as a player moves around, information about game objects (such as the buildings) will be sent to the player.

Likely, if there are players in the city, most buildings will probably be loaded into state 2 (partially loaded) and will stay that way. Suddeny the player picks up a torch and tries to set the building on fire, the building will be loaded its fuil state from the database into memory (fully loaded as in state 3), the fire from the players torch can now effects the buildings state.

One thing that will help with dynamic environments is the property system.

Basically it is about setting key, value for game objects state.

If you have an apple, you can set the following properties:

{type, fruit}

{fruit, apple}

{color, green}

{state, fresh}

If you have a blender, you can put a game object inside the blender, the blender will check for the properties, type, fruit, state and color  and produce "green apple juice". If the state property would have been rotten, the blender would have produced "bad apple juice".

Systems like this has been used for several years, NannyMud for example (http://www.lysator.liu.se/nanny)

I hope that answers your questions.