[self-interest] Re: An OO history

[Reinout Heeck]

Samantha Atkins wrote:
> 
> Reinout Heeck wrote:
> >
> > Ruby took the 'best' of a couple of languages among which Smalltalk, so
> > it had to 'fix' the Smalltalk syntax.
> >
> > Since the Byte article on Ruby I lost interest in trying it, it looks
> > broken to a Smalltalker, specifically the absence of named parameters
> > (keywords), the implied declaration of block parameters (if a method
> > takes a block as an artgument Ruby uses zero characters to indicate
> > that) and the dichotomy between blocks and closures all seem a step
> > backwards to me.
> >
> 
> You would reject a language over named parameters?  To each their own I
> guess but this seems like a weak criteria.

I wrote that that is _just one_ of the reasons so the answer to that is
no.
That said I happen to find named parameters a *great* instrument for
both readability and defect avoidance. Therefore I find it strange that
a language that claims to take the best of a couple of languages among
whitch Smalltalk  didn't adopt them.


> Blocks are not closures.

True, but blocks can be treated as a degenerate case of closures like
Smalltalk does, so the need for two different concepts and syntaxes
falls away.


> True closures don't exist in Smalltalk.

Huh? I guess you are mistaken here. As I understand it blocks in
Smalltalk are true closures.

If a Smalltalk method returns a block and that block references local
variables of that method, the block's variable references are bound to
the values of the variables as they were during the method invocation.

Why don't you consider that 'true' closures?


>  Smalltalk syntax leaves much to
> be desired.  Which is not surprising.

I disagree strongly, I find it one of the cleanest and easiest syntaxes
I encountered, yielding very readable and compact code with a very low
defect count on first composition. It is one of the few syntaxes that
can be learned *completely* within one or two hours.


>  It is surprising to see a
> language with rather different design and usage criteria criticized for
> not being enough like Smalltalk.

Since the author claims to have borrowed ideas from Smalltalk I find it
strange that some very good things got omitted/altered for the worse. 


If you feel like continuing this thread I suggest we find a more
appropriate forum for it.


Cheers!

Reinout
-------