Antwort: Mixed feelings
Kamil Kukura
kamk at volny.cz
Fri Jun 7 16:40:44 UTC 2002
Boris_Gaertner at msg.de wrote:
>
> Kamil wrote about his impressions and posed some questions.
> Here are remarks to some of his questions:
>
> >>For example, Object understands so many methods that I doubt it still
> >>stands for basic building stone.
>
> I agree that the Object protocol became huge over time. Methods like
> #beep are not really necessary, but they are convenient.
And it will be bigger by time. For example, existence of method
Object>>asHtmlDocumentForRequest: indicates that we can just expect more
and more selectors being added as productivity goes on. I was thinking
about idea of good protocol category description and then having
selective inheritance so that a class would inherit either all or some
(selected by protocol categories list) methods from superclass.
<snip>
> >>Or - LargeNegativeInteger is subclass of LargePositiveInteger -
> well......
>
> Well, here the classes have the responsibility to represent the sign.
> This is an implementation trick. (By the way: IBM Smalltalk uses a
> class LargeInteger to represent both positive and negative large
> integers)
That's clear, but that design. Having class LargeInteger and then two
subclasses of it (positive and negative integer) makes more sense to me.
<snip>
> >>Would it be possible to extend language so it could be
> >>written as something like #perform:*with:?
>
> I failed to fully understand this question.
I know what is #perform: I just meant to have written -
perform: aSymbol *with: #(firstObject, secondObject, thirdObject)
instad of perform: aSymbol with: firstObject with: secondObject with:
thirdObject
> All my interesting porjects are still unfinished, so I have
> nothing that I could give you. This may change in a few
> years, but for now I have nothing. (Until tomorrow, I will
> prepare the polynomials for you)
I read the polynomials example, thank you.
--
Kamil
More information about the Squeak-dev
mailing list
|