[MUD-Dev] Distributed virtual worlds (Was NEWS: Why Virtual Worlds are Designed By Newbies - No, Really (By R. Bartle))

[Sean Kelly]

On Tue, 4 Jan 2005, Tom Gordon wrote:
> Sean wrote:
>> On Mon, 29 Nov 2004, Yannick Jean wrote:

>>> I may be missing something here... If the codebase is just
>>> minimally data driven, transferring the book datas from one
>>> database to another is truly trivial.

>> Certainly.  But if you're talking about moving data between
>> different game engines then the data will likely be meaningless,
>> unless the import process involves some sort of translation
>> procedure.  Say you import a longsword from a D&D rulesystem to
>> one that uses weapon types as damage modifiers, and one which has
>> no built-in longsword template.  The weapon might be plausibly
>> effective in the fiction of the destination world, but making it
>> work with the engine is another issue entirely.  I'm sure that a
>> normalized data format could be agreed upon, but making it useful
>> would still require a large amount of cooperation.

> Of course, someone could always write a base protocol that
> describes how data can be moved from one environment to another,
> and let each developer create mappings based on the protocol,
> rather than having to actually have any co-operation.

But letting each developer create mappings *is* cooperation, isn't
it?  If I don't want to cooperate then I don't create any mappings
and the data is meaningless in my application.  Sure, it would be
trivial to allow a player to carry data around in its generic
format, but making that data usable in an environment requires
interpretation.

> What springs to mind immediately from the networking world is
> Shibboleth (http://shibboleth.internet2.edu/) and JANET-LIN.

> The point of mentioning these is not *what* they are (which is
> unrelated in reality), but more about the idea and concept behind
> it: Shibboleth was created independantly, and the people who make
> use of it map it to their own internal systems. It's a generic
> protocol, implemented as needed.

I think I'm missing something.  Aren't the two concepts synonymous?

> So, someone who publishes the data may have a 'longsword', but
> when it is described in the shared protocol while it may have a
> described name or type listed as 'longsword', this can be mapped
> to whatever, and the destination environment may not have a
> 'longsword' template - and may not need one, as it can map the
> incoming protocol stream to it's own internal systems.

But any way you cut it, the destination system has to do its own
mapping to make the data meaningful, which was my point.  Whether a
'longsword' is represented as a long pointy metal object or as
someting else entirely must be determined by the parameters of the
system.

Sean
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://kanga.nu/lists/listinfo/mud-dev