Pro / cons of the various kind of MMOs ?

[puggy]

Thanks for the post, very informative. Can I ask you some questions?


What if you use ajax though? not like flash, but you can have some dynamic content on screen. my game would be niche so not really lots of anims, so ajax would work just fine (I never tried the language though)


Why you think is the longest to make? assuming you use something like python or TGB which has all the network stuff built-in, I can just call php/python script on the server to make the logic of the game (remember my particular case is turn-based without direct interaction with other people apart trading, forming alliances, etc).

Also why you say it's most likely to make more money? that's the opposite many people are saying

p.s. and yes I agree with you about FB!

The ajax stuff is part of the 'ways around that'. I've seen some very nice site's out there which make use of things like that to create better looking and functioning websites, unfortuantly there's not that many (relativly speaking).

As for longest to make, yes you could use 3rd party stuff to cut down on the time, though it could restrict you on what you can do so getting something that works for what you want is important. Even mixing a client up with asp/php parts is possible (starpeace uses asp for parts of the game).

As for the money part, it's all about percieved value, a downloaded client based game has more percieved value than a browser based game. So it's easier charge people more, however it is harder to gain those users, so while you make more money per person, you will also have less people. whether you make more money overall one way or the other is hard to judge and depends on the quality of the game and the quality of your promotion.

tolik: yes, html/php/asp will use a lot more bandwidth, i know this very well as my mmo uses both, on average a user has about 700B/sec usage and the web parts are about 56k on average on the page plus images. The images do get cached by the user but the users can access page's every 10-20 seconds and the web parts arn't the most important part of the game. It's the extra html tags and rubbish that up's the web parts as the actual useful data is a very small part of that. Stripping the html stuff and about 1-2k per page is useful.

As for the app data, check out both the developer and application leaderboards user DAU (daily average users) and see how quickly number 1 zagna at 64.8million go's to number 40 aa at 0.5 million. It's a steep drop in only 40 developers from a list of 33230 developers. Thats not even taking into consideration the number of those which are silly wee apps that get high numbers of a month or so and then die.

[Hideo]

Well, well. I will have to reconsider... Some folks are indeed making MMOs alone. BTW, I didn't notice it was Jack. I thought it's another of those "boy wants to make a WoW clone" type of questions?

Nonetheless, I still think MMO is one of THE most demanding projects one can attempt. It requires so much dedication to maintain one. It's only for the most patient of us.

[Psychochild]

I'm a fairly infrequent poster, but saw a link to this thread from GameProducer.net. I ran the game Meridian 59 for about a decade. I've also advised a lot of indie MMO developers, including Over00 and his game Golemizer.

First, I'd advise making sure your scope is small for your first game. If you're interested in the technical bits of making your own MMO, start small. Over00 tried to do a big sandbox game for his first game, and while it is impressive, he points out that there's a lot that could have been improved. MMOs aren't like other games where you can start on version 2 right after making the first one. If you get people playing your game, you'll have to support it and keep those players happy; otherwise, people won't trust you enough to stick around for your second game.

I would also advise against trying to go at it completely by yourself. Yes, a single person (+ outsourcing) can make an MMO. But, you can do so much more by sharing the load, usually. Of course, you always have to deal with unreliable people unless you're paying salaries. Even if you are paying people, there's the chance they might take the money and just run.

Anyway, if anyone wants advice on building and running an MMO, feel free to visit my blog linked in my signature. Or, you can drop me an email. Sadly, I don't check these forums as often as I used to, so PMs on here might not be the best way.

At any rate, good luck to anyone who wants to do an MMO. It's probably one of the more extreme forms of indie game development you can do. It's also very high profile since generally you make your money by getting people to play over a length of time. Consider your options carefully!

[Over00]

I'd like to put focus on Brian's (Psychochild) post here.

While I was alone behind the keyboard while coding his support and advice were what get me through the end of this project.

In no way it is easy but it can be done. Having someone that kick your butt once in a while and bring you back to reality helps a lot.

The scope of the project is important to define. I started with a wild idea and somehow made it to reality though maybe not the way I would have hope. If I would get to start again the scope would be much more smaller, specially since I had no knowledge at all of thing like community management, decent marketing and ways to monetize the game.

If you're wondering of ways to properly monetize your game than I'd advice to drop Brian a mail about it. He's the one that can really help you on this topic.

[Jack Norton]

Yes, I made some experiment in last days and I realized there's no way I can make it alone I'm probably going to outsource the big programming part and concentrate on what I do best, the game design/marketing/management (that's a good excuse eh).
Also I don't want all my time sucked into a MMO because if doesn't do well, I'm probably going to make less than now. So will definitely make an attempt with a simple one

[puggy]

I wouldn't count out a WOW clone either. While doing the whole thing from scratch would be a really REALLY bad idea there are also tools out there to do it.

Realmcrafter ( http://www.realmcrafter.com/ ) is basically an mmorpg system where is just scripting, layout and models and there pro version looks very nice. Start off with a small world and build up and you could get a nice small community though the MMORPG market is saturated so don't expect big things.

infinity (http://www.infinity-universe.com/ ) looks like it has the potential to compete with eve though it has been in development for a very long time with only a few developers part time and a lot of contributions model wise.

the universal ( http://theuniversal.net/ ) is a nice mmo that been around a long time and has evolved a lot with only 2-3 coders, though it's more of a fun project rather than a money making game for the developer.

A client/server game can be done my small or even one man teams it's just a LOT harder to do and maintain.

To be honest, the best advice i could give is if your thinking of going this route is to split your time with the major MMO with some other project like a normal games or a smaller php or flash games, that way you can make some money along the way. You could even use the smaller MMO php or flash games to get some experince though stepping up to a server/client system is a big step.

[Jack Norton]

I actually bought realmcrafter a while ago (I think 2 years ago). But I never managed to use it, and doing a 3d MMORPG game is probably the worst thing I could do
(too much competition, lots of expenses for 3d models, etc)
I actually could probably spend 1 year working on the MMO with my recurring income, but I am not sure I want to be "chained" to a single game.
I mean with my normal games, once finished, I don't need to worry about them anymore. I see now that while MMO are cool and can make good money they also require LOT of time to mantain and/or a full team, and I'm not really sure that's what I want

[defanual]

I'd probably go with Flash (and php, ajax and whatever else with it) if I was doing something 2d.

3d, maybe this: http://indie.bigworldtech.com/index/index.php Seems like their tools and infrastructure would take out a lot of the leg-work http://www.bigworldtech.com/technology/index.php

[tolik]

tolik: yes, html/php/asp will use a lot more bandwidth, i know this very well as my mmo uses both, on average a user has about 700B/sec usage and the web parts are about 56k on average on the page plus images. The images do get cached by the user but the users can access page's every 10-20 seconds and the web parts arn't the most important part of the game. It's the extra html tags and rubbish that up's the web parts as the actual useful data is a very small part of that. Stripping the html stuff and about 1-2k per page is useful.

This is blasphemy as I've mentioned - AJAX is a solution and pointed to lordofultima.com.

As for the app data, check out both the developer and application leaderboards user DAU (daily average users) and see how quickly number 1 zagna at 64.8million go's to number 40 aa at 0.5 million. It's a steep drop in only 40 developers from a list of 33230 developers. Thats not even taking into consideration the number of those which are silly wee apps that get high numbers of a month or so and then die.

First of all, learn to spell Zynga. Then, define a point of interest - that's apps with >50k daily audience. Then, count those companies. Hint: Hundreds.

[tolik]

BigWorld is too big for small projects.
With Jack's Python knowledge, he could easily make a game with SmartFoxServer (ugly outdated AS, Python or Java) or even ElectroServer (Java, JavaScript or AS in that case).

[Jack Norton]

With Jack's Python knowledge, he could easily make a game with SmartFoxServer

That's interesting you are referring to:
"Embedded webserver supporting Java/Python servlets" feature of the PRO version ?
What exactly does, dynamic webpages with python scripting?

[tolik]

Check out few samples in the docs. You are able to script entire server game logic in Python. The client could be Flash, Shockwave, .NET, Unity, AJAX, etc.

I would still recommend to go with PHP, as SFS is realtime game oriented server.

[Jack Norton]

Well my problem isn't the server, I can code in php, but the client, since I can't code in ajax/flash at all. I could make a very static full php game but I think would be a bit boring

[tolik]

Well, take Push Button Engine framework or even flixel and start learning!

[Jack Norton]

Haha no, I am old and lazy
I managed to create a login/create new account interface with a renpy (python) client + php server in about 2h today, knowing nothing about network programming.
I can see though that doing a MMO is really a long project, so much stuff to check server-side to avoid cheating, etc.

While I am fairly confident I could go on with this client/server system on the technical side, I am not sure I'll have enough motivation / time / patience to finish it! I really admire everyone that finished a MMO, is really a big endeavour.

[zoombapup]

Apparently PBE was used in social city (big thing on facebook right now).

We're experimenting with flex + php but in all honesty any backend language will work as long as you have plans for scaling.

For my part, I'm considering appengine (googles thing) running either python or java as the authentication system monitor etc. Then running amazon EC2 servers (started on the fly based on traffic volumes as seen by appengine) to scale out as needed. The appengine code will simply be a load balancer (which has the added advantage of always being there, where as the amazon EC2 instances might not).

A lot of what I'm planning is procedurally generated and partly seeded from the facebook data of players (literally seeded, so for instance place names would be munged from the name of their friends).

All seems eminently do-able (I'm a network programmer of old) but the big issue for me is literally how to ensure it doesnt take man-years to maintain. So lots of issues in terms of economy balancing, making sure that there are elements worth sharing etc.

Different kind of game and design challenge and I'm definitely not doing it to jump on a bandwagon, but I see this as an interesting way to get a game live that does some of the things I'm interested in.

[tolik]

zoombapup has a plan!

(but I don't get appengine idea as you don't measure load by amount of traffic flowing through load balancer, but by the actual metrics such as cpu load and load average on the servers. plus there's native load balancing from ec2 or the simple way of proxying apaches through nginx frontend/balancer)

[zoombapup]

zoombapup has a plan!

(but I don't get appengine idea as you don't measure load by amount of traffic flowing through load balancer, but by the actual metrics such as cpu load and load average on the servers. plus there's native load balancing from ec2 or the simple way of proxying apaches through nginx frontend/balancer)

I'm not really talking that kind of load. I'm talking load as in player count per area. The game idea is all about exploration and trade, so I want there to be a certain amount of traffic in any geographical region. So to disperse the players at an even rate, the appengine will be the router between initial login and actual starting point for new players.

So each area of the game world, will be populated by the appengine, who will also generate the world on-the-fly. As new areas are explored, the appengine app will generate new content based on the first explorers facebook data. Essentially as someone pioneers a new route for trade, they get to name the route etc. Then the plan is to allow designers (i.e. me) to go in and add additional content (or generate it) on top of that base for more interesting world stuff (like special event kind of stuff).

The game is meant to be mostly async but with a bit of player interaction in that the trade prices for any given item fluctuate over and aggregated price (so you cant literally screw someone over by dumping one commodity in a place, but overall the price gets effected on aggregate).

Plus of course, appengine can manage the EC2 instances.

Now I'm still torn on where to store the data for each player. I guess the EBS block store is the right place for that, but I'm not entirely sure there. Lots of silly little details to work out there.

[zoombapup]

btw: for those interested, check out flex builder 4. It has a pretty nifty new set of interfaces for hitting up remote servers and doing calls on them. Handles http, json, xml, php etc. Very nice if you're not exactly a network database guru.

There's a 60 day trial. Might be fun to try and actually set that as a hard limit on the time to prototype your game!

[tolik]

Now I'm still torn on where to store the data for each player. I guess the EBS block store is the right place for that, but I'm not entirely sure there. Lots of silly little details to work out there.

EBS is more of an archiving solution (e.g. for stats data). You should store all the data in MySQL right on your EC2 machine(s).

[Nutter2000]

For our next game, we're looking at PBE with an ASP.net backend (possibly) running on VMWare ESX servers rather than the cloud.


However, for your first MMO, I would highly recommend sticking with what you know, Jack.
In your case, maybe an offline client with php/python backend?

It IS a massive undertaking.
Still you don't have to do a major league MMO, especially not on your first attempt.
Why not take one of your old games such as Universal Boxing Manager and create a small MMO out of that.
It's niche enough to attract interest without much competition (afaik) plus you already have a client pretty much done.


@zoombapup
I'm definitely looking at investing in Flash Builder 4, FlashDevelop just doesn't really cover everything enough with it's lack of design tools.

..and Visual Studio 2010....

Damn.. need to start saving

[Jack Norton]

Still you don't have to do a major league MMO, especially not on your first attempt.
Why not take one of your old games such as Universal Boxing Manager and create a small MMO out of that.
It's niche enough to attract interest without much competition (afaik) plus you already have a client pretty much done.

Yes I definitely DON'T want to do a big MMO. I made some tests and yesterday already finished a login/create new user system with python. So in next days I'm going to work on it, but as secondary project since I don't want to invest too much time on that.
I know probably doing a webgame would be x3 easier to promote, but in my specific case it would be x10 harder to code, so in the end I prefer to finish something with what I know and see how it goes.
I'm going to design it so that it will work as a singleplayer with extra features, so I can always sell it as normal game even if the MMO part doesn't go as I expected

[Nutter2000]

Yes I definitely DON'T want to do a big MMO. I made some tests and yesterday already finished a login/create new user system with python. So in next days I'm going to work on it, but as secondary project since I don't want to invest too much time on that.
I know probably doing a webgame would be x3 easier to promote, but in my specific case it would be x10 harder to code, so in the end I prefer to finish something with what I know and see how it goes.
I'm going to design it so that it will work as a singleplayer with extra features, so I can always sell it as normal game even if the MMO part doesn't go as I expected

Yeah, that's pretty much what I was trying to suggest

Play to your strengths and you'll have more chance of succeeding.
That's why I suggested UBM, it's something that you've already done client side, so you could create a simple webservice, online database of players, boxers for hire and matches, some sort of league system and of course modify the client side to include challanges, and talk to the server.
Just an example, pinches of salt and 2 cents (or denomination of your choice) apply

FWIW Cliffski's doing something similar at the moment with GSB.

Good luck for your next project anyway and don't be swayed by the naysayers

[zoombapup]

EBS is more of an archiving solution (e.g. for stats data). You should store all the data in MySQL right on your EC2 machine(s).

Hmm, it would make me a bit worried if the data was stored in MySQL on an EC2 instance given the instance can fail and drag the data with it. Mind you they do allow some sort of volume mounting on it though dont they..

My main issue is that the game isnt entirely async. If it was, then I'd do the MySQL db route and forget about it. Maybe I need to challenge the design issues there and see if it really is worth it.

[Game Producer]

Slightly derailing the conversation to a different direction, but I must add one thing (for anyone who ponders "if I can make MMO"): get this book called Designing Virtual Worlds.

It's a must read for anyone who even thinks of thinking of doing an MMO.

[tolik]

Slightly derailing the conversation to a different direction, but I must add one thing (for anyone who ponders "if I can make MMO"): get this book called Designing Virtual Worlds.

I would suggest to... AVOID ANY DESIGN BOOKS ON MMO.
Avoid old school advices from people who've worked on subscription-based MMO games. AVOID THEIR ADVICES. Find your own way.

[zoombapup]

I would suggest to... AVOID ANY DESIGN BOOKS ON MMO.
Avoid old school advices from people who've worked on subscription-based MMO games. AVOID THEIR ADVICES. Find your own way.

Actually, its a pretty good design oriented book. I doubt its got any advice that wouldnt work for a social online game.

The main thing everyone preaches is "metrics metrics metrics" which is fine. Web developers have been doing that forever. But it wont teach you how to make a great game, so much as give you insight into what your players like.

[Nexic]

I kind of agree with tolik on this. I read that particular book and I found a lot of the ideas incredibly outdated.

If I remember correctly one such piece of advice was to completely do away with any game mechanics that might 'annoy' the player, such as penalizing them for not performing some kind of character maintenance x times per day/hour. Ironically a lot of ultra successful browser games now use maintenance as a core game mechanic. As it turns out this is a great way to keep people playing, which ultimately leads to a lot more money.

Again, I may have remembered wrong, but I believe the book also said that allowing people to buy power for real money was a bad idea. However, doing this is a brilliant way to make money and most players now see this as a normal.

[zoombapup]

Just wanted to mention that Rob Zubeck did a great talk at GDC 2010 about creating scalable servers for social games. (Rob works at Zynga).

The main takeaways are:

Use flash + php + memcache + mysql

php takes care of most of the heavy lifting in terms of gameplay. You use memcache where possible to pull data rather than always hitting the database. You use mysql and replicate using data striping so that you can scale out the database traffic. Load balancers on the front end (between flash and php) allow you to hide the php servers and load balance across the system properly.

Of course, this only really works for async data where players arent in direction communication. He mentions scaling out similarly where you DO require that, but doesnt really go into any detail there. I'd imagine something like smartfoxserver or the usual java enterprise stuff would work there if you need that kind of persistance.

He also talks about trying to keep your databases from having to talk to each other.

I've still got a few issues, revolving around specific database access patterns (for instance he talks about accessing one row per webserver instance rather than trying to hit all rows from a single one, so scaling out row-wise rather than table wise).

Well worth a watch if you have access to GDCVault.

[zoombapup]

I kind of agree with tolik on this. I read that particular book and I found a lot of the ideas incredibly outdated.

If I remember correctly one such piece of advice was to completely do away with any game mechanics that might 'annoy' the player, such as penalizing them for not performing some kind of character maintenance x times per day/hour. Ironically a lot of ultra successful browser games now use maintenance as a core game mechanic. As it turns out this is a great way to keep people playing, which ultimately leads to a lot more money.

Again, I may have remembered wrong, but I believe the book also said that allowing people to buy power for real money was a bad idea. However, doing this is a brilliant way to make money and most players now see this as a normal.

Hmm, thats not what I take away from it. What I take away is that you should plan for players play styles changing over time. Accept that you are going to have some people who want to screw up the game for others etc.

Of course, if you have an async game where you never have players interact directly, then its moot. But I think the other stuff is worth considering even if ultimately you reject it.