🚀 go-pugleaf

I am hoping to spark a discussion about what the "Logo" of the 90s might
look like.  I recall hearing Seymour Papert about 15 years ago say that
Logo was designed by taking the best ideas in computer science at the
time (late 1960s) and "child engineering" them.  In particular, Lisp
heavily influenced the design.  As I recall Seymour said that a mistake
of the Logo group was not keeping up with developments in computer
science.  He mentioned actors and object-oriented programming.

Here we are in 1992.  There are lots of good new ideas about programming
coming from logic programming, concurrency, linear logic, visual
programming, and program animation among others.  I imagine that many
people reading this will be thinking, "Gosh, it's really hard for
computer scientists to understand the recent developments in many of
these areas, how can he expect this to be suitable for kids?"  Well, I
think that was probably the same reaction that the original designers of
Logo and Smalltalk received when they talked about their work.  Twenty
years ago it was a very radical stance to claim that a programming
language for kids should borrow heavily from such sophisticated
languages as Lisp and Simula (object-oriented programming).

I am aware of some attempts to move forward.  Object Logo tried to be
object-oriented while keeping Logo as a sublanguage.  Ten years ago at
Imperial College in London researchers like Kowalski and Ennals tried
teaching Prolog to kids.  Mitch Resnick's masters thesis (1988) is about
concurrent programming in an extension of Logo (and his more recent work
is to bring massively parallel SIMD computation (like the Connection
Machine) to children with *LOGO).

I sense, however, that most people on this list are pretty satisfied
with Logo as it is.  They want better graphics, use of windows, menus,
multiple turtles, etc. but think only in terms of incremental
improvements to the language and environment.

I see Logo (and Smalltalk 72) as very good first steps towards a goal of
giving children powerful, yet easy to learn and use, means of
programming computers.  To control multiple Lego motors and sensors, for
example, from a monolithic sequential programming language such as Logo
leads to very non-modular programs that scale poorly. I think that the
next "Logo" should be more powerful (based on concurrent objects) and
easier (based upon visual programming, program animation, and a simpler
underlying computation model).

Comments?

 -ken kahn