Re: argument about encapsulating data sublanguage

Thomas Gagne wrote:
> I admit to not getting the humor or the irony (I'll duck behind
> something now or "self thickSkin: #on").
>
> Many OO systems already have what Marshall describes above. To go a
> step further, Smalltalk doesn't even implement math natively. "x + y"
> sends a message to x telling it to run the '+' method with the argument
> 'y'. That's how Smalltalk can be taught o add apples and oranges

This is wrong. You send x and y messages to the Plus object. Then the Plus object asks authorization of the AlgebraManager to perform such an operation. Upon approval the Plus object delegates the job the lowly IncrementByOne object. Received on Thu Dec 28 2006 - 22:32:07 CET