Croquet Project

[tkr]

Hi,

I am just curios what you think about the croquet project (http://www.opencroquet.org). Do you think that this architectur is applicable in a game?

tkr

[Mr_Light]

they seem to sugjest such a thing: http://www.opencroquet.org/About_Croquet/screenshots/slideshow30.html

I have some resevations towards the consept, but I'll refrain from judging some I didn't know about 10 mins ago. 

[Jeff]

We jut had a talk on it inside of sun that OI missed...

But I have a strong feeling they dont adress any of the real issues that games face such as security and latency hiding.

Might be useful for a cooperative builder-MUD type environment where you have trusted users.  I doubt its useful beyond that.

But I could be wrong.

JK

Games published by our own members! Check 'em out!

[tkr]

I have some resevations towards the consept, but I'll refrain from judging some I didn't know about 10 mins ago. 

What are your reservations?

The croquet architecture in one maybe not meet the needs of games. But whats about the TeaTime approach inside of croquet. A peer-to-peer "scalable real time multi-user architecture" [http://www.opencroquet.org/Site%20PDFs/Croquet%20Collab%20Archi%202003.pdf] to communicate between objects sound for me interesting. Because, you don't have to leave the abstaction layer of java and do some selfmade binary communication beween your clients. The only issue I do see is security, because peers talk to each other. But maybe this is the only way to develop a MMOG and not have a supercomputer at the backend. To manage all the synchronization.
 

[Jeff]

MMO engines have been tried where the work is distributed to clients.

Its soudns great, use your user's compuerts to scale... but in the end it all comes back to security and the fact that there provably is no way to achieve it on the client.

So again, if you have a population if users you can gauarantee are cooperative and wont try to hack the game then its a fine approach, otherwise its a nightmare waiting to happen.