discussion: limits of objects

[Mike Prince]

On Wed, 7 Dec 1994, Francois-Rene Rideau wrote:

> So, as of inter-code communication, the *big* problem with our
> migrating objects (and we know that in a persistent, distributed,
> system, they will migrate a lot, and even the seldom migrating will
> migrate between backup/persistent store and RAM), is what are the
> limits of an object ?

We need to develop a way of measuring the strengths of an objects ties to 
others.  One controversial way to to ascertain this a priori.  You state 
A links to B,C,D, with each having a value.  With a bunch of objects 
having all these relationships, you could determine (very costily) who 
should be moved.

Another way is to do it on demand, not the most effective, but very 
easy.  Trying to spawn one too many agents?  Your toolbox gets bumped to 
a lightly loaded machine.  Trying to gobble up too much RAM,  again you 
get bumped.

An earlier opinion was to separate policy from mechanism.  This mandates 
that we must have a technique for gathering this relational and usage 
information, even if we don't use it from the start.  Someone else can 
use our data to establish the policy later.

Mike