[DRAFT] Co-located services

[Clint Byrum]

Excerpts from Gustavo Niemeyer's message of Fri Jul 29 08:03:51 -0700 2011:
> > The second risk is that this system is abused in the wild to escape
> > possible limitations of the current system with regard to machine
> > reuse. This could inadvertently allow for a new class of bad practice
> 
> Well put.  We need to enable multiple service units per machine before
> colocation.

This exchange makes me wonder if the focus is to enable use cases, or
to change peoples' behavior. The latter is an order of magnitude harder
than the former. I am totally up for a challenge, however I want to make
sure we all are, because up until now the focus has been on use cases.

If people find that their use case is served by "abusing" a model in a
way that works for them, we should *embrace* that. Words like "bad" and
"abuse", therefore, should not appear in specs in this context.