🚀 go-pugleaf

rslade@cue.bc.ca (Rob Slade) writes:
>One of the chief benefits of Logo, in my experience, is that the use of it
>somehow improves and directs programming "style", regardless of the prior
>experience with other languages.  I have not analysed this in detail, but
>I firmly believe that the programming "contructs" (and syntax) of Logo have
>a lot to do with it, including the absence of GOTO.

I hate GOTO, too, but actually it's only quite recently that there have been
Goto-less Logos.  Apple Logo and IBM Logo, for example, both had goto (it was
called GO).

The Logo version of goto is really pretty limited, since it only works within
a single procedure.  This is both good and bad; the good side is that you
can't write *really* awful code with Logo GO, but the bad side is that long-range
jumps are what one mostly wants goto for!  Things within a procedure can usually
be done just as well with some iterative mechanism.

On the other hand, Logo does have two goto-like constructs that can be used to
write spaghetti code if you try hard.  One is CATCH and THROW, which provide
a nonlocal exit facility.  The other is tail calling!  I've seen plenty of Logo
programs in which each procedure ended by calling another procedure.  In effect,
the procedures were used as code blocks linked by goto.  No procedure ever
returned to its caller.

So, I agree that leaving goto out of Logo is no big loss, but neither does it
ensure well-organized programs.