A possible future... - ryzom.org forum archive

An archived topic, merged from 22 captures made between 8 January 2007 and 21 June 2008. ryzom.org · Forum index · All captures

Author

Message

< Build A World General ~ A possible future...

Talad

Posted: Wed Jan 03, 2007 11:48 pm

Joined: 03 Jan 2007 Posts: 3

Hi,
I'm Luca, Director of PlaneShift project (www.planeshift.it). I'm quite sad that the ryzom initiative could not take off well, and this just adds up to the list of failed mmorpg projects we all have in front of our eyes.

We at PlaneShift are working very hard to provide a great experience to the players, while keeping the engine Open Source. Until today we have been very successfull in our organizational model and in our operations. We have 3 tech demos delivedered, 300.000 registered users and about 200 concurrent players always connected to our servers. It's not bad for a game that's not yet completed. We have a passionate group of fans, that built guilds and web sites about PlaneShift, and we are willing to expand the game as much as we can.

If you want to see a free mmorpg online, real, playable, enjoyable, if you want to dedicate some of your skills and time to game development, there is not a better place than PlaneShift team.

Please consider not creating another mmorpg, but joining our forces to have greater chances of success. Download the game, check the web site.

We are willing to select talented and committed people, improving our team with your ideas and help. The selection process is not easy, but I assure that it's worth working with great people, and I've very proud of the accomplishments of our current team. If you want to build something real, join us.

You can reach me at [email protected]

Hope to hear back from you,
Luca.

barometer

Posted: Wed Jan 03, 2007 11:56 pm

Joined: 28 Dec 2006 Posts: 31

That is one way to go and sounds like a good solution. The down side, it´s
aclosed evn.
What people here was talking about was a project in true Open Source spirit,
that is...everyone who wants to may contribute.

Talad

Posted: Thu Jan 04, 2007 12:37 am

Joined: 03 Jan 2007 Posts: 3

I don't want to start a long debate, because you have books and books on the topic, but my view is that Open Source on art simply doesn't work well for artists and for project evolution. You can read more on it in our license pages.

barometer

Posted: Thu Jan 04, 2007 12:45 am

Joined: 28 Dec 2006 Posts: 31

Yes, a can understand a closed project as well... benefits of it and why your
team want's to work in that specific way.
In my opinion, let's say ( some one on the learning stage of 3D modeling )
may contribute to a large project, even if the material that can be used, then
there can be ideas someone else can go further with.

_________________
Project Angela
http://rpg-gamerz.com/smf/index.php

valentin

Posted: Thu Jan 04, 2007 12:59 am

Joined: 07 Dec 2006 Posts: 23

Talad wrote:

Hi,
I'm Luca, Director of PlaneShift project (www.planeshift.it). I'm quite sad that the ryzom initiative could not take off well, and this just adds up to the list of failed mmorpg projects we all have in front of our eyes.

We are willing to select talented and committed people, improving our team with your ideas and help. The selection process is not easy, but I assure that it's worth working with great people, and I've very proud of the accomplishments of our current team. If you want to build something real, join us.
Luca.

hi luca and thanks for passing by.

shure planeshift does exist, i play it.

as barometer says our goal is to allow any team to 'plug a shard', so many teams can live and make games.

if something in between full art disclosure and nothing able to compile from source is possible we really should have a talk.

what we are looking for is a packaged working gpl demo for teams to build upon. we can do and/or finance the packaging from a cvs repository or such.

valentin

Cayle

Posted: Thu Jan 04, 2007 9:57 am

Joined: 20 Dec 2006 Posts: 82 Location: Germany

Hi Luca,

I understand your reasons for not releasing the art assets; I understand that artists are more protective of their work than programmers are. I understand that some of those things are what makes Planeshift Planeshift rather than Generic World X. Lastly, I also understand that some things simply require out of pocket expense; sometimes a considerable out of pocket expense and it is simply rude to ask for those things with no compensation. The last thing I want to say is "i want you to do this because it greatly benefits me and i don't care if it ruins you", which is often what OSS/GPL mantras can be.

I second valentin’s notion. Sourceforge is littered with dead persistent world projects. People start out starry eyed, wanting to build “teh uber” and they loose momentum after a couple of months. Where I see my thoughts coinciding with his is in the creation of a reference world and assets. A reference world; a simple one that people could use as a learning example and a set of community art assets could allow small teams to bootstrap their visions. It would lower the barriers to entry for would be world builders. The worlds created would be mostly poorly designed ones that look alike because they use only the communal assets, but there would be some gems among them.

My ideal would be the creation of something along the lines of the Neverwinter Nights persistent world community. That game (referring to both nwn1 and nwn2) is by no means intended to be a MMO/PW platform, but people put it to use in that way. The main reasons are that there is an engine, art assets, etc either packaged with the game or available from its very active community. A small group can build a persistent world on nwn. Another reason so many nwn pw teams stick to that platform is that as of January 2007, no real viable alternative exists for anyone who does not want to start from a blank slate. Bear in mind that "a small group can build a PW" is not the same as a "click here to make my dream". It just makes it attainable for small groups willing to work hard for a year or more.

Let me ask openly, would you be opposed to such a reference world being built on the CrystalSpace/Planeshift engine?

Cheers,
Dave

_________________
The Ruleset Modeling Language - A project for making it easier to design virtual worlds.

Han

Posted: Fri Jan 05, 2007 12:53 am

Joined: 21 Dec 2006 Posts: 34

Thanks Luca,

I have tried and enjoyed Planeshift also - its good stuff all around. I'm not sure what will happen here - I would hope some of the more experienced Ryzom (programmers ect) guys would take the helm and get us moving toward something useful. I lead usually, but have no want to do it here. It is not my place to do so. Just trying to figure which way the wind blows before setting my sails :)

Hannon

Lurker

Posted: Fri Jan 05, 2007 11:54 am

Joined: 27 Dec 2006 Posts: 2

Hi "Talad",

I also have to side with Valentin and "Cayle", my interest is in putting my 2 c ents into a game.

Playing with 2 or 3D graphics is fine, but unless you can put your creation up somewhere for other to critique, continuing becomes difficult, one needs the feedback and encouragement to keep going a become an expert. I lack the skill to build en entire enviroment, so plugging in my 2 cents somewhere, even if temporarely is very appealing.

I'd also like to see other aspect evolve, like the dynamics of combat, only casters seem to have some degree of this at this point, Cayle has some intersting ideas in this area (project Angelica). I'd like to work on a more natural human interface take the example on Wii. I've just seen a videos, of Wii interfaces to WoW
http://www.youtube.com/watch?v=ev9ZhYiUs4g
http://www.youtube.com/watch?v=jkX1oQsvNe0
neat!
Imaging adding basic character movement, combat slashing (right to left, top to down for example. One could probabls go even further (kicking, ducking ... ) ...

Yes, I too can understand the need for a closed project, but why not have small separate sub-projects, like "barometer" suggest, to flush out the good ideas. In essence these would be parallel Planeshifts (no pun intended :P )

Daniel

Vengeance

Posted: Fri Jan 05, 2007 1:27 pm

Joined: 05 Jan 2007 Posts: 7

Hi,

I am one of the primary programmers on the PlaneShift project and I've worked on it for 5 years or so. Despite a licensing structure designed to have people not reuse the PS artwork in other games, the engine itself is fully open source and GPL. You can download the code from SF anytime you want and there is enough "sample" art (also freely licensed, as is required by SF) there for you to not only build the engine but to run the game and run around inside its virtual world.

If a group wants to use the PS engine and make a shardable game with more open content, I would support that effort 100% and help them try to be successful. Talad's view is that he really just wants people helping with PS, which I understand, but imho independent efforts like this will not hurt PS and will only help the engine become more advanced, etc. This is the beauty of FOSS.

Talad

Posted: Fri Jan 05, 2007 1:34 pm

Joined: 03 Jan 2007 Posts: 3

Ok, it seems to me that you are more on the lines of "let's build a base engine, so other people will be able to make a free mmorpg". The fact is that when you will have the engine in place, you have to hope for someone to start from there and actually make a game.

I go in another direction "let's deliver a playable, enjoyable game, with the engine open so people can learn from it and build on top".

My objective is to deliver a game, fully functional. The world is full of unused engines.

Anyway, like Vengeance said, PS is already a usable engine you can take and make a mmorpg with.

valentin

Posted: Sat Jan 06, 2007 1:06 pm

Joined: 07 Dec 2006 Posts: 23

Vengeance wrote:

hi,
planeshift is shurely appealing, but we were somehow turned down by this thread about making a new mmo from planeshift.

http://hydlaa.com/smf/index.php?topic=23915.0

then we got turned down a bit more by the poor packaging as seen in the compiling instructions.

http://planeshift.cvs.sourceforge.net/*checkout*/planeshift/planeshift/docs/compiling.html

so we got the idea (wrongly i hope) that planeshift would not help us releasing a 'plug and play shard' for any team to use under gpl licensing. and it's not a good idea to start a developpement against it's base team.

we shurely do not want any kind of forking or dilution of energy.

i wrote to talad about un upcoming meeting probably next week so maybe the best way to go is that you attend this meeting also (by phone or physically in paris, france) so we can openly meet each others and discuss our and others will.

in two words our base plan is a common platform enabling a team to 'run a shard' from scratch. this means client and server, but also account creation, enough 3d art and recipies to have an appealing working game and examples to build upon, and all the nice packaging to make it really usable without internal knoweldge at first.

we envision such platform would be appealing enough so lots of teams use the same graphic chain, avatar description, object description, world description, avatar interaction process.

this would not weaken existing projects while allowing newcomers to play (wich is our point).

we know of two candidat platform existing today that could lead to our point in a short timeframe (less than a year), delta3d.org and crystalspace/planeshift.

valentin

Vengeance

Posted: Sat Jan 06, 2007 5:29 pm

Joined: 05 Jan 2007 Posts: 7

Valentin,

I can understand from that thread why you might be discouraged by it. :) However, on that forum we get a constant stream of 14 yr olds who want to run a copy of PS or make a new game with it themselves. We tend to discourage the fanbois from attempting it because they underestimate the level of effort by 2 or 3 orders of magnitude of what it really takes to build an MMORPG, whether you have a base engine or not. That is the thrust of that thread.

As AryHann posted there, she used the PS engine in her Univ thesis to make a virtual archaelogical dig so that paraplegic kids could take virtual field trips there as a group and take a tour, just like with the real site in Sweden. While she got a lot of advice and help from the PS team, she did it all basically on her own, so even a one-person team can make something meaningful if they are dedicated. The group here seems fairly professional and fairly serious, so I have no doubt you could be successful with the PS engine also.

As far as the difficulty of building the game goes, I don't know what to tell you. It is certainly complex but there are numerous other OSS projects that also are a part of PS, so "building" is actually building 5 or 6 different projects in sequence. It could be nicer but we've never had one of those guys on the team who finds it endlessly fascinating to make the builds perfect (you know who you are). :) If we can get it to work and it is "relatively" painless, that is good enough for most of the team. In fact, since we get a steady stream of people wanting to volunteer (1/day or more), we tend to use the build process as a way of filtering out the people who don't have the skills or who aren't serious. We do help everyone who builds it, though, and even have a dedicated irc channel called #planeshift-build to share knowledge about it.

The data and content in cvs is minimal because the devs focus on the code and only put in the data they need to a) test their code, and b) illustrate examples of what is required by the Rules and Settings teams. However, there are dozens of items, a dozen or so npc's, sample quests, spells, faction groups, guilds, crafting sequences, etc. If someone wanted to make the "out of the box" cvs content richer and more interesting for people to try out, the PS team would welcome the assistance.

Now I will "sell" you a bit on our engine, if I may. :) It is still a work-in-progress, obviously, but it is fairly mature in many ways. Any major feature in WoW, for example, can pretty much already be done in PS. Features would include:

It also does all this quite efficiently. We currently average about 200-250 concurrent players on our demo server, 24x7, with about 300 active npcs, and the server process uses around 8% CPU on a machine smaller than the laptop on which I'm writing this post. The network averages <1KBps per player. The code base isn't 100% immaculate but it is fairly well structured and easy to understand if you are a good C++ person used to very large projects.

I wasn't aware of Delta3d and it certainly looks cool. It looks very client-focused though, and not really an MMO engine, so I guess it depends on what your goals are.

Anyway, this post is probably long enough by now. :) Thanks.

Induane

Posted: Sat Jan 06, 2007 8:07 pm

Joined: 03 Jan 2007 Posts: 8

Making a new game with PS's engine would require a good bit of knowledge and work but it is possible. I've slowly began a site on documenting PlaneShift. I host my own build guides which are much easier than the official ones imho, and I include instructions for how to test new maps ingame, which is a key step in re-using the engine. I'm only one person and not a PS dev myself so I don't have all the info but everything I've figured out is here:

http://vaalnor.mine.nu/psdoc

it may be of use.

_________________
- brant watson

Grom

Posted: Sun Jan 07, 2007 5:07 am

Joined: 18 Dec 2006 Posts: 11 Location: Poland

Planeshift make one, but very big mistake - He use CrystalSpace engine. Crystal space is fat and ugly. This engine maybe have many features but any programmer dont use 50% of this features. Its to big for big project. CL must have many ram, good graphic card etc. If Planeshift will be growing, will be have big requiments too. I think, if Planeshift programmers dont modyfi game code and dont minimalize game rquiments, this game will die :( Sad but true.

Sorry for my English etc. Maybe someone undestand 10% of my text :D

Cayle

Posted: Sun Jan 07, 2007 7:40 am

Joined: 20 Dec 2006 Posts: 82 Location: Germany

Vengeance wrote:

Valentin,
Now I will "sell" you a bit on our engine, if I may. :) It is still a work-in-progress, obviously, but it is fairly mature in many ways. Any major feature in WoW, for example, can pretty much already be done in PS.

It also does all this quite efficiently. We currently average about 200-250 concurrent players on our demo server, 24x7, with about 300 active npcs, and the server process uses around 8% CPU on a machine smaller than the laptop on which I'm writing this post. The network averages <1KBps per player. The code base isn't 100% immaculate but it is fairly well structured and easy to understand if you are a good C++ person used to very large projects.

Hi Vengance, I'm happy to hear that my first impression on your team's willingness to work with others was wrong. I also have to say the low resource useage of the Planeshift server is a point in its favor over using Delta 3D, which to my knowledge has not yet been used in an attempt to make a persistent world.

Vengeance wrote:

Features would include:

This is an impressive list and covers most of the basic services that players would expect on a fantasy persistent world. I do have a couple of questions however. How deep does the planeshift specific implementation of the rules go? This is important as every little group that uses any reference platform is going to want to do its own thing. For example, if someone wants to substitute a different crafting system, is he looking at Python and XML only, or c++?

I'm also curious about how action queue management and the AI of mobs works. This is where any ruleset would "meet the metal" so to speak. How is it laid out?

_________________
The Ruleset Modeling Language - A project for making it easier to design virtual worlds.

valentin

Posted: Sun Jan 07, 2007 11:58 am

Joined: 07 Dec 2006 Posts: 23

seems i was wrong and i'm delighted :)

with a documentation effort and planeshift team help, making a reference platform will be a breeze (well not as easy but from what happened here in a few days it seems lots of help would come on such a project).

pleasure,

valentin

Vengeance

Posted: Sun Jan 07, 2007 3:18 pm

Joined: 05 Jan 2007 Posts: 7

I don't really know how to answer the question of how PlaneShift-specific the rules are. The engine assumes a certain structure of when RPG rules are applied--such as calculating dmg when you swing your sword... how much HP does the mob lose? But the formula for that calculation is scripted outside the C++ in an xml file.

By putting it in xml, we free our Rules people from having to rebuild PS every time they want to tweak and try a different formula. It also turns the RPG formulas into "data" instead of "source code", which means it is not subject to the GPL, which means we don't have to publish our formulas publicly. This keeps PlaneShift "special" and keeps sophisticated players from reading our formulas and knowing specifically how to optimize their characters for our formulas. The data in cvs has example formulas but not the "real" ones.

This xml would enable a different project to create their own combat rules very easily as long as the basic structure of how it works agrees with what the PS engine is expecting. So for example, if a player jumps off of something, we calculate their fall damage with this:

<script name="Calculate Fall Damage">
exit = (rnd(1) > 0.5);
Damage = if(FallHeight>1,pow(FallHeight, 1.8) * 0.8, 0);
</script>

As long as you make a script with that name, with FallHeight as an input and Damage as an output, you can make your Fall damage calculation totally different. You would probably want to do that because how many HP each player has in your game would probably be different from ours.

If you wanted to add additional RPG formulas or change inputs and outputs, C++ changes would be required but it isn't a big deal to do that. We would just need to code them in a way that a superset of all our requirements was hooked by the engine.

This RPG script mini-language is one of 5 or 6 different domain-specific scripting systems within the PS engine. They are all fully documented and compiled to very efficient form on startup of the server. The client receives almost none of this data because, in an OSS project, hacked clients with compromised cheats coded in is a real problem. PS was designed from the ground-up to keep source level cheats to a minimum--although more work needs to be done in this area.

I'm not sure what you mean by action queue management. The AI of mobs is a separate and large topic so I'll make another post on that after I eat some breakfast here. :)

valentin

Posted: Mon Jan 08, 2007 12:42 am

Joined: 07 Dec 2006 Posts: 23

k, lots of teams are going torque or nwn because of ease.

if we provide a way for those teams to go 'planeshift compatible' wouldn't it help everybody ?

what i would envision is;
- a creative common art/rules repository directly usable in crystalspace/planeshift
- a packaged planeshift snapshot using these art/rules as a demo shard
- a friendlier reception of wannabe teams enabled because they can go mess what they want on their side using the above tools

a first try could look like a theme park demoing the usable arts/rules, it should be quite easy to set up compared to a full project. (?)

valentin

Vengeance

Posted: Mon Jan 08, 2007 7:42 pm

Joined: 05 Jan 2007 Posts: 7

I think it would be fairly easy to produce such a thing if you can come up with the art assets you need. The models and textures are really the most critical part and are the hardest to get produced.

If you already have those or can come up with some, I'm sure myself and the entire PS team would be happy to help you make it work. You could have such a demo shard up within a month, or less.

Vengeance

Posted: Mon Jan 08, 2007 7:59 pm

Joined: 05 Jan 2007 Posts: 7

I forgot about posting on the AI topic yesterday, so I'll do it now...

Our design goal from the beginning has been to support >1000 concurrent players on a single server box. Depending on how skinny or how advanced your AI is, AI can be quite memory and CPU-intensive, which is in conflict with the 1000 user goal. We also wanted to support both "dumb" AI for rats and bats, as well as support "smart" AI for advanced boss mobs in case some AI doctoral candidate decided to use our project for his/her thesis.

Thus, we split out the NPC management task into its own process. It connects to the server much like a regular player client, but it uses a slightly different networking setup which is optimized for the necessary Entire World view as opposed to the players, who need high-precision in a small subset of the world. Since it is connected via sockets to the server, it can run on its own box (or boxes). Since this separate process is doing all the AI, movement and scripting of the npc's, all the server is really doing is accepting its commands about what its characters are doing, much like the server does for a player client also--except that the npc client is sending commands for 300 characters instead of 1! :) This is why you will often hear the npc client referred to as the "SuperClient".

Also, since multiple npc clients can connect to a single server, we can have a "farm" of AI boxen depending on how advanced the AI gets, and each one could have a different implementation of how it does its AI, as long as it follows the npc client network protocol the server is expecting.

This is a long way of saying that the AI scripting capabilities in PS today are not the only ones which could be supported. I would actually love to see a whole new superclient made using more advanced techniques.

What we have today is scriptable npcs which can move, attack, chase, run away and so forth based on "events" such as being attacked, proximity, faction or time of day. It is more rudimentary than I would like, and the team is currently working on some new amazing things with it, but it serves the purpose for now. While many in the OSS gaming world seem to have dreams of super-smart mobs, our philosophy is slightly different in that we try to keep in mind that the mobs exist to be killed by the players--they are nothing more than cannon fodder. If we had really smart Genetic Algorithm trained rats, they would learn to avoid the players and no one would ever find a rat. :) And it turns out that dumber mobs take less CPU, so we like that aspect also.

We have detailed Word docs on our current superclient implementation, npcclient, and how its scripting works, etc. which we can share with you when you are ready to get into more detail.

Hope that helps,