Re: service stereotype thingies

[Elliot Lee <sopwith redhat com>]

On Fri, 20 Oct 2000, Michael Meeks wrote:

>         A service offers more characterization power than oaf. Oaf can
> only characterize objects that come out of factories.

Eh, what do factories have to do with it? Whether or not a factory is in
use is opaque to the client making use of OAF.

> It can't speak about the services provided by a moniker resolution [
> although in fact resolving a moniker against a service makes good
> sense to me ].

Monikers will find object impls by using OAF - the former proposal was to
add the idea of a service to OAF queries.

> > Another way to solve this is just to make sure that each "service"
> > (component, object impl, whatever) has a primary interface associated
> > with it, and if an object implements that primary interface it also
> > implements all the other interfaces associated with that "service". It
> > seems like an interface for each service will be needed anyways in
> > many cases.
> 
>         Providing a new primary interface per service ( as was suggested
> earlier ) that does nothing except act as a tag you can QI for was one

The interface would provide some actual functionality if there was some
needed (e.g. serve the role of both the Cell service and the XCell
interface). Doing a QI shouldn't be necessary AFAICS, unless one wants to
verify that a given object provides a given service (which doesn't seem
something one normally does).

-- Elliot
DEAR IRS, I know you got my unsubscribe request. please stop sending me
all this junk mail and asking me to work half the year for you.