Monday, January 7, 2008

Decisions, Decisions

One of the more promising ideas that I have come up with is working on a MUD like game that is completely web based. This has some advantages in that it wouldn't require much artwork, and I could spend most of my time working on the engine and mechanics of the game. It would also allow me to experiment with some different technologies that I have been wanting to work with.

There is still a small piece of me that would like to do a graphical game, good artist or not. I'm also unsure how much interest there would be in a new MUD today.

Whatever decision I make, I have come across some materials that have an interesting back story and play system. It is called Enoch: Rplaying in the Second Gensis. Some things will have to be modified a bit to facilitate a MUD environment, but I think it looks very promising.

Decisions, decisions, decisions...


mike barton said...

Good question -- the MUD world has all but died over the last 5 years. Mailing lists that used to be hard to keep up with now sit dormant.

OTOH, Kingdom of Loathing and games like Kings of Chaos have gotten wicked popular over the same time period, despite having interfaces far less rich than what I'd envisioned.

I figure if you make something that's fun, has a lot of personality, has a low bar of entry, facilitates socialization (even, if not especially, the killing-each-other kind), and market it properly, it'll be successful.

Um, I could expound on all of that, but this is probably already too lengthy.

Donovan said...

You should do a web-based MUD. The advantage of doing a web-based MUD is that the required techniques (AJAX and Comet) are entering a renaissance period, so it's easy to figure out how to do it correctly.

In my "spare time" a couple of years ago I wrote some pieces of one. Here is the svn checkout command if you want to take a look:

svn co

It doesn't even run on the latest eventlet/mulib but it has some good ideas in there I'd like to resurrect.

Now that eventlet has the process module, it should be trivial to take the main part of this,, and do what I originally intended to do with it, which is to run untrusted code in sandbox processes. The idea would be to run one process per "agent", where the agent scripts are run, and one process per "room". Inside these processes each object is an actor, which means it gets it's own "process" in the form of a coroutine. Objects would communicate via http, which is also much easier now that eventlet.httpc and eventlet.jsonhttp exist.

Chuck said...

Hey Donovan,

That sounds cool... I'll have to check that out soon when time permits me to start playing with it again.


Chuck said...


I got a chance to check it out, and there is some cool stuff in there. It has also inspired me with some ideas. Thanks!