[DGD] just out of curiosity

[Noah Gibbs]

The distributed game problem (I won't specifically say MUD) is going to look completely different in 10 years, so it's just as well you don't spend 10 years solving it ;-)

At this point, there's no such thing as a MUD that won't fit on a single server.  Before solving the distributed MUD problem, you have to figure out what problem it is.  That affects a *lot* of how you approach it.

The problem won't be more hand-built content than a single database server will hold.  Database servers are simply too big to have that problem on any reasonable hand-buildable scale.

If the "distributed game problem" is a problem with having multiple *people* run them so you need multiple servers, you first have to solve the rules problem -- how do you keep other people from sabotaging you?  How do you avoid various kinds of well-intentioned cheating?

Nobody has ever come close to solving *that* problem, and it's not a technical one, so technical solutions fall flat.  No really, if the other guy your MUD is federated with starts giving away Giant Glowing Swords of Awesome more easily than he realized, now you've both got a really hard problem.  And that kind of problem can crop up weekly.

As far as the actual interaction model, though, servers are freaking powerful now.  Use database transactions.  Use simple in-memory objects, possibly with regular statedumps.  Use anything you like.  There is no good solution to this problem because no real MUD *has* this problem.  Perhaps there was a brief period in the MUD-size/server-power curve where somebody did, but not for very long.  This wasn't a particularly credible worry in 2002, either.  In 2012, it's ludicrous.

This would be a bit more of a problem with, say, World of Warcraft.  However, they don't have problems with complex code interacting at a distance with fiddly timing constraints because they do very little that uses complex code for interactions.

Now, if you first invent a really interesting interaction method that can be used in a variety of games and requires this problem to be solved, great.  Figure out *what* problem you're solving, though, or your solution won't work well.

Yes, it's possible to come up with thought exercises where you could strain the computing resources of a mid-range modern server.  You know why people don't fix those problems?  Those thought exercises are 1) optimizable and 2) not fun.

Shentino's case is different in the sense that he plans to compute across huge scales.  The thing is, he's assuming content somehow magically pops in, basically in infinite amount.  If you do the math on how many people would need to build and for how long, it would take so long that you wouldn't be swamping a then-modern server any more.


________________________________
 From: Felix A. Croes <felix at dworkin.nl>
To: dgd at dworkin.nl 
Sent: Friday, September 14, 2012 6:34 AM
Subject: Re: [DGD] just out of curiosity
 
Ragnar Lonn <prl at gatorhole.se> wrote:

>[...]
> I'm a little bit skeptical about the feasibility to make existing 
> MUDs/mudlibs "evolve" into graphical ones. I think that any MUD/mudlib 
> that gets graphics added to it at a later stage, will have text and 
> graphics that are mismatched. The whole thing will look awkward. To get 
> a really good graphical MMO, you have to think about the graphics from 
> day 1, not when you already have a system with functionality and 
> processes that are all about text content.

I meant the evolution of the mud community, rather than individual muds.
Once one text mud starts with a javascript client, others could pick it
up and start to improve it.


> If you're aiming for the cloud, wouldn't automatic state distribution 
> across the network be the absolute killer feature for DGD?
> (One can always try...)

Efficient automatic state distribution for a mud is a hard, unsolved
problem.  I may not be smart enough to solve it, but if it can be done,
it can be done in LPC.

As for killer features, making Hydra took me 10 years.  DGD cannot
survive as a text mud server for another 10 years (or however long it
would take me to properly solve the distributed mud problem).

Regards,
Felix Croes
___________________________________________
https://mail.dworkin.nl/mailman/listinfo/dgd