Interfaces

Alejandro F. Reimondo alereimondo at sugarweb.com
Wed Feb 4 08:47:41 UTC 1998


Matthijs van Kempen wrote:

AR> Do you know the time you invert in interfaces (defining, mantaining) ?
MvK>NO
AR> Are they usefull for you?... or for your compiler. :-|
AR> Are they for people or only for the compiler (or the component technology)?
MvK>Not direcly, but they are not that different from the categories we 
already have. 
Interfaces consume much more time than categories, because they have more information.
Categories are "objects" that help humans, they are not used by object message sending mechanism. They only help in seen the messages (by category).

MvK>I agree that objects can be more powerful, but I think a class 
hierarchy alone could not offer the ease of use needed for rapidly 
building apps.
Yes, it do. I have written apps rapidly during last 10 years.
How much rapidly can be with interfaces?
Most of the time "the bugs" in Smalltalk does not occur in protocol mismatch (provably because you live with the objects all the time), and when it occurs, is because your objects are evolving/mutating, and it is a great thing in an environment, if it can be immediate (at the speed of thinking).

MvK>I'm thinking more along the lines of Morphs encapsulated with all  
the code they need. Make them distributable and you have something 
magnificent.
MvK>You could even transfrom a Bookmorph into a real compound document 
that could travel over the net.
MvK>Say something?!
Very interesting.
Affect them with dynamic interfaces, but not to all objects of our world. :-)
If interfaces are normal objects, objects can negotiate dynamically the way to talk!
and what they need to work properly.
	(container support: plugIn interface)
	ifTrue: [ container meet: plugIn ].

Cheers,
Ale.

Alejandro F. Reimondo
Feel free to visit Smalltalk User Group of Argentina (SUGAR)
at http://www.sugarWeb.com





More information about the Squeak-dev mailing list