Re: But is it Object-Oriented?
Isaac Gouy <[email protected]>
| Newsgroups | gmane.comp.lang.nice.general |
|---|---|
| Message-ID | <[email protected]> |
'many practicing programmers used to single-dispatching object-oriented languages complain that multi-methods just don't feel object-oriented.' > Their later solution was to revert to the generic function approach > and solve encapsulation issues at the module level. IMO, for a > multimethod language this makes the most sense. I only quoted that paper as evidence that there was nothing new about asking "But is it Object-Oriented?" with reference to multi-method languages. > In Java/Eiffel/C++ there is the idea that an ADT is implemented by a > class. In ML, Haskel, and languages with generic functions, it would > be more appropriate to say that an ADT is implemented by a module, > where a module is a collection of types and methods defined on those > types. "A class was defined as both a structural system component - a module - and a type. In contrast, a package is only a module." Meyer's viewpoint - refering to Ada packages (of course, imperative langauges such as Modula, Ada, Oberon-2 implement ADT with modules.) When we talk of interfaces in Nice, we should be thinking about the features that are public in a package - not about a type. As you've said, the Java "interfaces" idea makes no sense when the unit of encapsulation is module, not class. Perhaps we should look to the Dylan module and library approach? http://www.gwydiondylan.org/books/dpg/db_201.html __________________________________ Do you Yahoo!? Yahoo! Finance: Get your refund fast by filing online. http://taxes.yahoo.com/filing.html ------------------------------------------------------- SF.Net is sponsored by: Speed Start Your Linux Apps Now. Build and deploy apps & Web services for Linux with a free DVD software kit from IBM. Click Now! http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click