Hi Marc, On Sat, 24 Feb 2001, Marc Lehmann wrote: > I'd like to remind people that corba is not the only way to go, as > there is also dcop and especially MCOP (which was designed for > realtime and multimedia applications). > > http://space.twc.de/~stefan/kde/arts-mcop-doc/ Very interesting, I was not aware of MCOP. > (the author couldn't get corba to work reliably (timing constraints) > so he choose to implement something suited for data transfer). I read through the MCOP docs, and saw very little of substance there to suggest why the author could not get CORBA to work reliably, or quickly, or indeed what the concrete advantages of MCOP were over CORBA. His point about 'bloat' was perhaps fair on the feature front, but there seems little reason not to use a sensible subset of the CORBA features and yet retain compatibility - something he does not do. Consequently I do not believe that MCOP's sacrifice of features, standardization, language bindings, compatibility, features and code re-use actualy gives sufficient / any concrete benefit for it to be worth using. But perhaps someone can further enlighten me. As for selecting by name, I'm waiting for someone to invent ACop before I switch :-) Regards, Michael. -- mmeeks@xxxxxxx <><, Pseudo Engineer, itinerant idiot
