[MUD-Dev] Dynamic Descriptions

[Chris Gray]

[Nathan Y:]
:To interject a new topic of discussion...
:
:Some of you have markup languages... I have one myself, but I am seriously
:thinking of replacing it. What I am concerned with is this - I want to be
:able to switch from running a text game to a graphical game of whatever
:sort without difficulty. To this end, I no longer pass any real text
:around... instead, I create information objects, which can key stored text
:(out of a database) or stored graphics, stored hyperlinks, etc. What is
:transmitted to the client is negotiated at login, and the server daemon
:(not the mud) resolves, for example, text from info objects to pass to a
:raw telnet client. What I am wondering is, do any of you have anything
:remotely similar to this, and if you do, are you willing to discuss
:technical details of it?

Not me; mine is all manual programming, down at the bottom - somewhere
in the scenario, code has to ask what kind of output the current client
can handle, and send that kind of output to it.

Where would your "info objects" live? If the whole thing has to be
transmitted to your "server daemon" (I just call 'em "agents"), then
you have then transmitted more info than if you just had the server
itself make the decision. Can your server daemons directly access the
info objects via some kind of shared memory interface? Doing so is a
bit dangerous, since they can presumeably be modified at any time.

Wouldn't you gain the same modularity, but without efficiency and
locking problems, by just using threads within the main server process
for the server daemons?

--
Chris Gray   cg at ami-cg.GraySage.Edmonton.AB.CA