[MUD-Dev] Object Models

[John Buehler]

Chris Gray writes:

> I've only been skimming the emails on this subject, but this bit caught
> my eye. I believe that a system like a MUD shouldn't be "compiled". By
> that I mean that it shouldn't, at its top levels, be fixed at compilation
> time as to what attributes and operations exist. This does *not* mean
> that I don't believe in strong typing. I do. My MUD language is strongly
> typed. With one exception: an object need not have a given attribute or
> operation available. Code can check, at run-time, for the presence of
> those things, and so make reasonable default assumptions. So, the test
> for being an "edge" can be made at run time. It need only be made in those
> places that require an "edge" thing to work with. The big advantage to
> doing this at runtime is that new attributes and operations can be added
> to an already running system, and, with proper checks and defaults, can
> work reliably with all things in the system, even given that those things
> have no knowledge that they are being checked for this new operation that
> they've never heard of.
>
> I think this kind of flexibility is quite important for long-running
> on-line systems like MUDs. Allowing them does not require throwing away
> strong typing at compile time. It just needs a simple escape mechanism
> by which checks can be done at run-time for the presence/absence of
> the attributes and operations involved. See the AmigaMUD web pages on
> my site for much more detail on what I mean.

Yes, others have been jumping on the dynamic reconfiguration issue.  I'll
concede that we didn't get around to working that out in our research team,
but we were getting there.  We had hallway conversations about it and I
believe that dynamic reconfiguration and upgrades are completely tractable
with a component system.

Anytime someone talks about runtime checks and such, I'll immediately raise
a red flag.  It's just rope around your neck.

JB


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