Re: Beginner's Language?

[Jochen Schmidt <jsc@dataheaven.de>]

FM wrote:

> Andreas Krennmair <a.krennmair@aon.at> wrote:
> 
>>The problem is that Lisp is in no way similar to a natural language. And
>>natural language can be understood easier than some functional,
>>theoretical stuff with a unique concept like Lisp.
> 
> You've got at least three assumptions there, none of which
> I agree with. I don't see why a programming language
> should be similar to a natural language (the only sense in
> which they are both languages is that they have some form
> of grammar) and it takes quite a lot of arrogance to say
> that a programming language is harder to understand than a
> natural language or that the imperative paradigm is closer
> to the natural way of describing things for humans than
> the functional paradigm.
> 
> Also, Lisp is hardly some functional thereotical stuff
> with a unique concept.
> 
> 
>>Don't get me wrong, Lisp is a great
>>language, and I personally find it quite exciting as four year Pascal and
>>two year C/C++ programmer, but it's hardly usable for teaching, because it
>>is so unique (well, you _could_ to functional programming in C, but it
>>would be considered bad style).
> 
> What is *so* unique? Lisp is a label applied to a broad
> family of languages, not a single language. That alone
> takes away the uniqueness. Furthermore, the functional
> paradigm is hardly confined to the Lisp family.
> 
> 
>>I currently go to an Austrian school where mostly programming/CS/software
>>engineering-related things are taught (it's for 5 years, and you start
>>at an age of 14),
> 
> Ouch. This is probably a bad idea...
> 
> 
>>> > Now look at it. Could a neophite with literally *zero*
>>> > experience in programming computers read it and stand a chance of
>>> > understanding what it does or how it does it?
>>>  
>>>  yes, yes
>>for my comment, see above. Lisp has hardly anything in common with natural
>>languages. Or could you express "My girlfriend's name is Suzy and her age
>>is 17" as readable as this in Lisp:
>>$girlfriend{'name'} = "Suzy";
>>$girlfriend{'age'} = 17;
> 
> Except that this is completely wrong. Your perl statements
> are imperative, not descriptive or even declarative. At
> least most languages in the Lisp family will let you write
> something close to:
> 
> (with (she (my girlfriend))
>   (be (of name she) "Suzy")
>   (be (of age she) 17))

(setf (gethash "name" *girlfriend*) "Suzy"
       (gethash "age" *girlfirend*) 17)

Is more like the equivalent to the Perl code.

but I would do:

(defclass girlfriend ()
 ((name :accessor girlfriend-name :initarg :name)
  (age :accessor girlfirend-age :initarg :age)))

and then

(make-instance 'girlfriend :name "Suzy" :age 17)

Which is IMHO much more natural than the Perlcode.

> That's aside from the fact that you chose an arbitrary
> sentence that isn't at all representative of natural
> languages. For a better exercise, take a few paragraphs
> from a book or better yet, record a phone conversation or
> radio broadcasts and come up with a formal notation that
> could be used to capture their meaning.
> 
> 
> 
>>Of course, this is not the best example, since Perl has a syntax that is
>>not the optimum for beginners.
> 
> Perl's syntax is the closest thing I've seen to a natural
> syntax in a programming language.

?? Is that meant as a joke or what????

Regards,
Jochen