Re: Beginner's Language?

[danfm@dartmouth.edu (FM)]

David Starner <dvdeug@x8b4e53cd.dhcp.okstate.edu> wrote:
>On Wed, 09 May 2001 18:01:10 -0500, Sashank Varma <sashank.varma@vanderbilt.edu> wrote:
>> can you provide some evidence that:
>> 
>> (1) any standard general-purpose programming languages resemble natural
>>     language at all, that there is some metric of natural language-ness
>>     by which they can be ordered, and that wirth-style languages score
>>     higer on this metric than lisp?
>>
>> (2) the degree to which a programming language resembles natural
>>     language predicts the ease with which it will be learned?
>
>Tell me what
>	!bbz#,^az@bz,
>does.

Depends on the language. Give me the language spec or a
reference implementation.


>Then tell me what
>	
>	OPEN FILE ARGUMENT 2 AS 2
>	BEGIN LOOP
>		LOAD FILE ARGUMENT 1 INTO BUFFER
>		COPY BUFFER TO 2
>	FINISH LOOP
>does. 

Same here.


>If I came upon the first, I would give up. The second, anyone can
>probably get some idea of basically what it's doing. For a new 
>student, knowing that they aren't totally lost can be a huge
>confidence builder.

It can also spawn a huge number of unwarranted assumptions
that will plague the process of learning. They will have
to learn to deal with the rigors of a formal grammar at
one point or another. The problem with the above grammar,
as far as I can infer from the code segment, is that it's
terribly complicated if you try to specify any substantial
subset of it.


>As for lisp itself, just from the name, what does defun or cadar
>do? What do FUNCTION, PROCEDURE and BEGIN do?

None of them does anything remotely close to what anyone
would be able to infer from one's knowledge of English
alone


>I've seen 2 + 2 since
>the first days of school. Why (+ 2 2)? Learning a new vocabulary,
>a new equation syntax is not what you want to be teaching when you
>have so much else to teach.

What equation syntax? Lisp doesn't have equations as part
of the language. Or algebra, for that matter. I'm rather
used to writing ab instead of a * b and how would you
support that? Allow only single-character identifiers?


>> (if you thinking i'm picking nits, note that i've granted you the
>> implicit assumption that there is a singular quality shared by all
>> natural languages, a chomsky-esque universal grammar perhaps, that i
>> could have also questioned.)
>
>That's a theoritical assumption. How about "The vast majority of the
>people who are going to be learning computer programming know an 
>Indo-European language, Russian, Chinese or Japanese."

I don't know how many of the above languages you can
speak, but there are enormous differences among them.


Dan.

-- 
What an author likes to write most is his signature on the back of a cheque.
                -- Brendan Francis