Re: A working document
Herman Bruyninckx <[email protected]>
| Newsgroups | gmane.science.robotics.orocos.user |
|---|---|
| Message-ID | <Pine.LNX.4.33.0112281036380.22078-100000@pc25-213.mech.kuleuven.ac.be> |
On Fri, 28 Dec 2001, Cezary Zielinski wrote: [... ``component image'' vs ``component interface''...] > > Why use another term when the whole computer science world uses > > ``interface'' as the term to represent the view that a component/class > > offers to the outside world? ``Interface'' can be used with both > > components and classes; it covers both data encapsulation as > > ``methods''. > > To me an interface is something between two faces - a kind of a go-between. > An image is something that is internal to a component. It is a data > structure ineternaly representing inside a component A another component B. > It is a representation (obviously simplified) of component B inside > component A. This data structure has to get its contents from outside the > component A. It gets this contents from component B through some kind of an > interface between them. > Thus two different names are needed. > This local representation of the interface of another component is called a ``proxy'' in computer science. The proxy is one of the best-documented software patterns around, so I would think _very_ hard before replacing its name by something else :-) Herman Herman