Re: INT: Worlds and Generic Services

Billy Tanksley (tanksley@owl.csusm.edu)
Mon, 19 Jun 1995 13:00:32 -0700 (PDT)


On Mon, 19 Jun 1995, Rainer Blome wrote:

> Chris wrote:

> > -Where are objects when not in a world?

> Objects should always be in a world (a context), even if it's the
> hardware (it's a hard world ;-).

That's how I implemented my object system in Forth. However, there is
something to be said for Classes-- essentially, they are objects that
have no context. Different systems differ on further details.

> > -Multiuser worlds ... would this result in pointless control fights?

> We can use the physical properties to break any ties - if two users want to
> manipulate an object, some force might give feedback to the users
> indicating to them that they are not the only one trying to use an object.

Some objects will be single-user; others will be multi-user. Perhaps the
application will only request a 'usable representative' of a class, in
which case you could use any object of that type that's in reach and not
locked down, or create a new one.

> Don't take me too literal here - no need for force-feedback input devices
> (they do exist), just any kind of feedback will do.

Right.

> Interfaces project and the rest of the Tunes project is unclear to me. I
> propose we define the domain of the Interfaces project as:

The interfaces project is entirely seperate from the rest on the Tunes
project.

> "The issues that pertain to the (virtually) perceived ("physical")
> properties of the objects that the user perceives when dealing with the
> system."

Skip the virtual; some of the things that are doing the percieving are
not humans.

> Rainer

-Billy