is squeak really object oriented ?
jan ziak
ziakjan at host.sk
Wed May 28 14:35:14 UTC 2003
On Fri, 23 May 2003 12:22:23 -0400, Joshua 'Schwa' Gargus wrote
> On Fri, May 23, 2003 at 11:49:19AM +0200, jan ziak wrote:
> > On Thu, 22 May 2003 23:19:54 -0400, Joshua 'Schwa' Gargus wrote
> >
> > a non-textual representation of a number two could be simply "two dots on
a
> > screen".
>
> That's natural.
>
> > and a representation of 324325435 can be a visual list of its digits
> > (digits are in the "dot notation"), thus 324325435 can be represented non-
> > textually. you can admit that i replaced the notion of a textual digit
with
> > the notion of some "dotty" symbols which stand for each textual digit.
>
> Whoa. I can't admit that. I can admit that you replaced one textual
> representation of digits with another textual representation. Are
> you really arguing that it would be better to have 9 dotty symbols
> and a 0-symbol instead of the symbols we are used to, and to compose
> larger numbers in exactly the same way with the new symbols? That
> seems like a very difficult position to defend.
>
i think you agree that the arabic notation and the "dotty" notation are
interchangable.
but i want to mention, i had not on my mind to discuss about representations
of numbers, i wanted to ask whether i always have to manipulate text composed
of LETTERS. the letters (as you will surely write back for their definition)
are defined as the signs which i have on my keyboard.
> (BTW, what is the non-textual representation of zero?)
zero is a symbol just like any other symbol, but your BTW question has
presumed that i should really depict a zero as nothing, so i do not want ot
answer your BTW question.
>
> > but the point is that representing ANY number on a screen in
> > non-textual form is indeed possible.
>
> As soon as you are talking about symbols that do not represent a
> number itself, but rather digits that must be interpreted together
> to find out what the actual number is, you are back to text.
>
> If you want 1000000 dots to represent the number 1000000, then I
> admit that you can represent any number on a screen (subject to
> screen and retina resolution limitations). However, you have not
> escaped text in any meaningful way by simply replacing numeric
> digits with dotty-symbols.
>
i have never written that i want to represent 1e6 by 1e6 dots...this was your
idea.
i agree that i haven't espaced it in your sense, but i have espaced it in my
sense because i don't have to manipulate text of english letters.
> > i think that we should adopt the opinion that every object can
potentially
> > have several representations (as several people in this list have already
> > mentioned) and take this opinion as common knowledge which need not to be
> > mentioned in the list. this is why i did not mention that objects can
have
> > several representations - it seems obvious to me.
>
> Ok. Then what do you mean by working with the object itself, and not
> a representation?
>
> >
> > > What does a collection of numbers look like? What about a
> > > collection of 324325435 numbers?
> > >
> > > > --- cut ---
> > > > the squeak system answers: "no it wouldn't. you must make a textual
> > > > representation of your object and ...".
> > >
> > > Yes, that would be nice. Maybe someday we'll figure out how to do
> > > it well. If you have some concrete ideas about how to implement such
> > > behavior, I'd love to hear them.
> > >
> > > In the mean time, Squeak has the really nice feature that a lot of
> > > basic objects such as Points, Arrays, Numbers, Symbols, Strings, and
> > > probably more print themselves out in the same syntax as used to
> > > create them. So, if you have an ordered collection in a workspace,
> > > you can highlight 'myOrderedCollection asArray', print it to get
> > > something like #(1 2 3 4 5 6.66), and cut'n'paste the already
> > > highlighted expression into your browser window.
> > >
> >
> > in some cases, it is crutial to perform operations with THE object and
not
> > with some "conversion" of it.
>
> If you mean that pasting the textual representation of the array
> into your code gives you a different array with the same numbers,
> then I agree. You might have an array with non-numeric objects
> that can't be pasted in this way.
>
> I'm not sure that this is what you meant, since the distinctions between
> the object itself, a representation of the object, and now a "conversion"
> of the object are not clear to me.
>
> Would you agree that you can never interact with the object itself, only
> with a representation, but that the textual representation is inadequate
> for your needs? That would be a good starting point, one that I can
> agree with.
>
i agree.
> > > Also, as others have mentioned, eToys and MorphicWrappers are worth
> > > looking into. I'm not sure if Self has been mentioned in these
> > > threads, but it might be to your liking as well.
> > >
> > > I'm not sure what you mean by 'work with objects not with their
> > > names'. As others have noted elsewhere, not matter whether the
> > > object appears as a textural name, a visual object, or a sound,
> > > they're all just ways of interacting with an object, not the object
> > > itself. Do you mean that you want to interact with visual
representations
> > > of objects?
> >
> > i had textual names on my mind.
> >
> > a position of an object on a screen is in fact its "name" in the context
of
> > the screen. i had not such "names" on my mind.
> >
>
> I do not understand your answer. Could you please restate it?
let me reformulate the basic problem: it would be nice to include a reference
to an object into a source code, so that i do not have to explicitly refer to
it by textual name. if you like, i would like to drag and drop that object
into the place where i use it.
i have looked at math-morphs, and i must say that it is not what i expected.
when i start typing "on air" then, again, i just type text (on my keyboard)
in order to refer to objects - it would by nice to mix the text with
references to objects i see or i find.
jz.
>
> Joshua
>
More information about the Squeak-dev
mailing list
|