IslandsMUD; source code released!
 

IslandsMUD is my entry in the 48 hour game making competition.

jplur.no-ip.org 5057

*passwords are currently saved in plaintext, please use something non-important*

( requires python 2.6 and )

It was really fun to make this under a time constraint. The source code is uncommented and a bit rough, I'm currently refactoring the base code and will post the new source when it is in a more usable state.

Comments / critiques would be great!

Re: IslandsMUD; python code base written in 48 hours
 

Don't have a real critique, but a cool idea! Were you familiar with Twisted before this LD?

Re: IslandsMUD; python code base written in 48 hours
 

Answered you online, but for posterities sake:

I've used twisted for a simple 3D TCP game, it's really easy but also a quite a large library. Thanks for suggesting MiniBoa, I'm tempted to switch over.

So I'm planning on making an actual MUD with this code, currently refactoring the base before I start adding new features.

Does anybody see a downside to using python pickle to save persistant objects? Originally I assumed I would have to use a database, but pickling my game objects to file is so simple.

Re: IslandsMUD; python code base written in 48 hours
 

It is simple, but the downside is dealing with version issues between, say, refactored classes and objects persisted from a previous version. You might want to look into just pickling dictionaries (or reading them from some other format, xml, yaml, etc.), and using that data instead.

Re: IslandsMUD; python code base written in 48 hours
 

There are no problems with using pickle for object serialization when you pickle the object's __dict__ (as IDE said). For AoA, I created a persistence mixin for objects whose dictionaries I want to pickle -- the mixin basically has 'save' and 'load' methods, along with some other class metadata that tells the base persistence class where to store the object and what to store...

As for a database, I've utilized data storage using Python's SQLITE interface for a while now, and quite frankly, storing the pickled dictionary in a column, along with desired data in other columns (IE: any object data you want as reportable/searchable; primary keys, etc) works very well. It's fast. It's easy. It just made sense.

Good Luck!

Re: IslandsMUD; python code base written in 48 hours
 

Great advice!

@ ArchPrime, I'll probably follow your lead and use sql so i can separate often looked up data into columns. (right now when you log in, it actually loads the player object before it can check the password.

@ Ide, yeah looks like it will be simpler to store player statistics in a dictionary. I'll still pickle the players inventory, since it's recursive it facilitates things like container objects.

Re: IslandsMUD; new source release!
 

Re: IslandsMUD; python code base written in 48 hours
 

Ah wow! I really like the look of that map. Are you thinking of going more "Roguelike"? Curious, as you mention taking your current experiment no further as a "room based MUD".

For further ideas that go beyond ASCII, I invite you to check out Archons of Avenshar. Granted, we've written a custom client, but the graphical minimap is amazing.

Re: IslandsMUD; python code base written in 48 hours
 

I second the suggestion of AoA (ArchPrime has really made a good game and client there, check it out). There are also some multiplayer roguelikes out there, Twilight MMORL comes to mind, plus third-party extensions to some of the popular RLs like Angband. There are others too that have been in development -- check this out




I think it'd be worthwhile to somehow keep the flavor of a mud along with the RL aspects though.

Re: IslandsMUD; python code base written in 48 hours
 

Ah, more great references to check out ;)

Yeah MAngbad (multiplayer-angband) is pretty great. Actually one of my other projects is a realtime 3D interpretation of angband:

I'll definitely be keeping some of the MUD flavor to the game. A lot of commands that were based on rooms will have to be funneled through a 'check nearby entities' function.

I'm also going to stick with Telnet for some time, it makes it easy for people to test it out. Once I know where this is going I'll develop a custom client.

Re: IslandsMUD; python code base written in 48 hours
 

haha, that's awesome!