Re: LOGO-L> Numbers or Words?

[tlbelding@htcomp.net (Tony Belding)]

On 26-Jun-97, Mike Doyle wrote:

>Saw the discussion on keeping data types distinct.

>I find comparison of Logo with other computer languages somewhat
>perplexing.

Maybe a bit of history is in order...

The story of high-level languages has always been one of trying to hide the
machine implementation from the user.  The ideal has always been that the
user shouldn't need to explicitly allocate storage for an integer, float,
string, etc.  The internal representation in the computer shouldn't concern
him, since it's only a distraction from developing his algorithms.
Instead, the language should contain facilities to transparently convert
data as needed.  That is the philosophy of languages as diverse as Logo,
REXX, and CanDo script, all designed for easy usage by people who are not
professional programmers.

Unfortunately, those conversions require CPU cycles.  They reduce the
execution speed of the language.  So, typeless languages have never made
much headway in the realm of professional or commercial software
development.  Those continue to rely upon typed languages such as C.  For
the sake of performance, they hire professional programmers who can be
expected to deal with such picky matters of explicit datatypes.


>As a teacher I use pencil and paper media which mix a great variety of
>'data types' before children's eyes. One of the  things they learn is the
>differentiation of these types, as a mental process. Logo models human
>though processes in this respect. Logo does not demand of young children
>an explicit understanding of data types before they write procedures. Logo
>itself differentiates where necessary.

Yep.  That's true not only for children.  LISP does the same thing, and
that's about as serious as computer languages get.  Don't forget that Logo
is a very high-powered language derived from LISP!  It's not designed only
for children.


>Which leads me to my view of the role of Logo. I do not see Logo as a
>precursor to programming in other languages.

Neither do I.  For one thing, it's not realistic to assume everyone will
become a professional C or C++ coder.  For another thing, Logo is more
powerful that most other languages.  The idea of "graduating" from Logo to
something bigger and better is backwards.  There is nothing bigger and
better than Logo, with the possible exceptions of Scheme and XLisp, or
maybe Smalltalk.  (I'm writing from ignorrance here.  I know Smalltalk only
by reputation, as the origin of object-oriented programming.)

As more powerful hardware becomes available, very high level languages such
as Logo only become more viable for a wider range of uses.  So, the trend
of using lower-level languages for the sake of efficiency has already begun
to reverse.  Witness the replacement of C by C++ and Java, and the almost
complete disappearance of assembler and Forth.


>I look to Logo being developed to reflect more effectively the true
>capabilities of the PC. At present it is stuck in an intermediate phase.

I'm not sure what capabilities you are referring to.  Currently I'm looking
at how to add some multi-media primitives to Berkeley Logo, such as simple
GUI requesters, load images, play sound samples, etc.  Is that what you
have in mind?

--
   Tony Belding
   http://hamilton.htcomp.net/tbelding/

---------------------------------------------------------------
Please post messages to the Logo forum to logo-l@gsn.org.  Mail
questions about the list administration to logofdn@gsn.org.  To
unsubscribe send    unsubscribe logo-l    to majordomo@gsn.org.