i'm not sure if anybody already thought of this, but i thought it would be a good idea: get a bunch of game developers who write the server-side code for network games to develop a specification and write up a RFC for standard gaming protocols for each different type of game. for example, there might be a protocol for 32x32 tile based rpg games, while there may be another protocol for 3d football games. the specification and requirements will be based on the most commonly used functions that these types of games require. and just like other RFCs, there will be mandatory requirements and optional functionality. the idea is kind of similar to the jakarta tomcat project (http://jakarta.apache.org/tomcat) as well as many others. there is a servlet container specification (established by sun microsystems?) that companies who choose to implement a servlet container must follow, meaning it must be capable of certain things as well as respond to things a certain way. just in case you don't know, the basic job of a servlet container is to run java server pages (JSP) and java servlets. tomcat is an open source reference implementation model of the servlet specification. companies can take this model and improve on it to add extra functionality. for example IBM websphere and BEA weblogic both meet the servelt container specifications, plus more. this allows other businesses to just use these server products to develop web client applications in java relatively quickly. now if the gaming industry had similar products, companies could just purchase game server software or use an open-source model with their game maybe even improve on it, that way they don't have to worry about most of the server and network code and can just focus on client gaming applications. client apps use this protocol, but each game would respond to different messages in their own unique way. just like how many companies use the same web server and protocol (HTTP), but their web applications differ dramatically. this benefits both game companies in speeding game development as well as other companies who want to sell gaming server implementations. anyway, i was in a rush writing this because i have to go, but if anything is unclear or you don't understand something i wrote, just ask and i'll try to do a better job of explaining it. just wanted to see what everybody thought of this idea. thanks. [edited by - _chasers on February 28, 2003 3:05:05 PM]

Trying to establish a protocol for this is unnecessary and impractical for the following reasons.

Standards slow development. A protocol to govern java web pages is fine, as you want all web pages to be standard. You also want all browsers to work in the same way so that everyone can access all the information on the net without having to run different browsers or install multiple plug-ins that all do almost the same thing. The downside is that it is very time consuming to get a standard agreed and once it is agreed there is less incentive for companies to try to develop something that goes beyond the standard.

Game development is an arms race. If one developer has more robust network code than their competitor their game will be seen to be better and will sell more. For this reason there is the incentive for them to keep refining their development. Also games are constantly evolving so you can't say that a particular game will always have a particular set of requirements.

Lastly it would be virtually impossible to get developers to sit down together long enough to agree a standard.

Of course the above does not mean that developers should always develop their own network code. I just don't think a "standard" is required. There are already companies out there selling or developing networking middleware for game development and I expect that more and more developers will adopt the use of "off the shelf" networking in the same way that 3D engines have been accepted.

Dan Marchant
Obscure Productions

[edited by - obscure on March 4, 2003 3:53:04 AM]