Suppose there is a hierarchy of refinements: the first message would
say "there is an object", the second would say "there is a large
object", the third would say "there is a large fridge", the fourth
could add information about colour and style, and so forth. This
gives the user the ability to stop getting more information at any
step in the process, either by the equivalent of clicking on the
spinning Mosaic world, or by having pre-set defaults in the viewer.
This could also apply to the world info: the first message gives the
dimensionality of the world and whether it's bounded, the second
could give more precise information about its size, and so forth.
With respect to position, the first message would say that an
object is at the lower left, an object is behind it, and another
object is at the upper right; and positions could be refined as
well. An example:
3D bounded NESWC:
object SE1, SE2, NW, C
link E, W, C
means there is a room bounded in all four compass directions and
in front (C for center) that contains four objects: one at the bottom
right, another behind it (but visible), one at the top left, and
one in the center. There are links to other places on the left,
the right, and dead ahead.
object SE1:
size large
name fridge
action open, close, get, put
says that the first object at the bottom right is large, it's called
a fridge (which can create an image if the name is locally available),
and that four actions can be performed on it (with constraints about
their orders following later).
I'm sure there are better ways to write these things; I haven't
been involved in the construction of MUDs, but experiences there
must be able to help.
------------------------------------------------------------------------------
Chris.Holt@newcastle.ac.uk ftp://tuda.ncl.ac.uk/pub/local/ncmh1/nameplate.html
------------------------------------------------------------------------------
I had rather be a dog and bay the moon, / Than such a cybernaut.