Re: [Maxima-discuss] LispTree: Draw Maxima objects as trees (using Lisp)

[Stavros M. <mac...@gm...>]

By the way, the original *Lisp 1.5 Programmer's Manual
<https://www.softwarepreservation.org/projects/LISP/book/LISP%201.5%20Programmers%20Manual.pdf>*(1962)
shows how lists are represented in Lisp using box diagrams starting on p.
37. Page 38 gives examples of common subexpressions (which Maxima allows)
and circular lists (which Maxima does not support). And page 41 shows an
in-place algorithm using *rplaca/rplacd*.

On Wed, Oct 30, 2024 at 2:30 PM Stavros Macrakis <mac...@gm...> wrote:

> When I was learning Lisp -- a very long time ago -- I found diagrams like
> this useful for two purposes:
>
> *1. Understanding structures with shared substructures.*
>
> Maxima does use substructure sharing, but it has very few destructive
> operations.
> So a beginner at Maxima (especially if they don't have a Lisp background)
> might have trouble understanding this:
>
> l1: [a,b,c]$
> l2: [d,e,f]$
> lll: append(l1,[l1],[l2],l2)
>     =>  [a, b, c, [a, b, c], [d, e, f], d, e, f]
>
>
> l1[2]:bbb$
> l2[2]:eee$
> l1 => [a, bbb, c]
> l2 => [d, eee, f]
> lll => [a, b, c, [a, bbb, c], [d, eee, f], d, eee, f]
>
> But it isn't necessary to explain this with box
> <https://www.gnu.org/software/emacs/manual/html_node/elisp/Box-Diagrams.html>
> diagrams
> <https://www.gnu.org/software/emacs/manual/html_node/elisp/Box-Diagrams.html> or
> tree diagrams. It's enough to note that *append* shares its final
> argument's representation with the result of *append*. In any case, to
> diagram this, you need graph-structured (not tree-structured) displays,
> which you don't support.
>
>
> 2. *Working out clever in-place algorithms*
>
> ... such as *nreverse*, which uses *rplacd* (See Steele's paper
> <https://bitsavers.trailing-edge.com/pdf/mit/ai/aim/AIM-587.pdf> for an
> example which is even more complicated, because he is working with cdr-coded
> data <https://en.wikipedia.org/wiki/CDR_coding>.). Using box diagrams for
> this can be helpful. But Maxima has nothing like *rplacd*.
>
>
>
> On Mon, Oct 28, 2024 at 8:42 PM Eduardo Ochs <edu...@gm...>
> wrote:
>
>> Hi Barton,
>>
>> you are right - the conversion from "Maxima objects" to "Maxima trees"
>> is done in Maxima, by the functions in lisptree.mac whose names
>> start with "lisptree0". Try:
>>
>>   load("/tmp/lisptree/lisptree.mac");
>>   block([inflag:true],  lisptree(a/b));
>>   block([inflag:false], lisptree(a/b));
>>   block([inflag:true],  lisptree0(a/b));  /* [*,a,[^,b,-1]] */
>>   block([inflag:false], lisptree0(a/b));  /* [/,a,b] */
>>
>> The function `lisptree0_apatom' calls `args', and if I remember
>> correctly it is `args' that chooses between the representations a/b
>> and a*b^-1 depending on the value of `inflag'... anyway, I just saw
>> that it would be useful to have functions that generate trees like
>> this one
>>
>>   MTIMES__.
>>   |       |
>>   $A      MEXPT__.
>>           |      |
>>           $B     -1
>>
>> from Maxima objects like a/b, i.e.,
>>
>>   ((MTIMES SIMP) $A ((MEXPT SIMP) $B -1))
>>
>> Here is a prototype. Run this in a REPL:
>>
>>   load("/tmp/lisptree/lisptree.mac");
>>   block([inflag:true],  lisptree(a/b));
>>   block([inflag:false], lisptree(a/b));
>>   block([inflag:true],  format0(a/b));
>>   block([inflag:false], format0(a/b));
>>   o : a/b;
>>   to_lisp();
>>     #$o$
>>
>>     (defun newsimplify (o)
>>       (if (and (consp o) (consp (car o)))
>>           `(,(caar o) ,@(map 'list #'newsimplify (cdr o)))
>>           o))
>>     (defun $new2dtree_ (o)
>>       (lisptree::toplain-lines (lisptree::lispytree (newsimplify o))))
>>     (defun $new2dtree (o)
>>       (format nil "~%~a" ($new2dtree_ o)))
>>
>>     ($new2dtree_ #$o$)
>>     ($new2dtree  #$o$)
>>     (to-maxima)
>>
>>   new2dtree(o);
>>
>> Here the output of its last line is this 2D tree:
>>
>>   MTIMES__.
>>   |       |
>>   $A      MEXPT__.
>>           |      |
>>           $B     -1
>>
>> Cheers, thanks for testing, more in other messages, etc,
>>   Eduardo
>>
>>
>> On Mon, 28 Oct 2024 at 13:32, Eduardo Ochs <edu...@gm...> wrote:
>>
>>> Hi Barton!
>>>
>>> I forgot to test it with inflag! I will do that very soon...
>>>
>>> I don't know how to use WxMaxima for non-trivial things. Anyone has
>>> any idea of how to make it print multiline strings using a monospaced
>>> font?
>>>
>>>   Cheers!
>>>     Eduardo
>>>
>>>
>>> On Mon, 28 Oct 2024 at 12:13, Barton Willis <wi...@un...> wrote:
>>>
>>>> Thanks for your contribution.
>>>>
>>>> Question: Am I correct that lisptree displays the internal structure
>>>> when 'inflag' is true and otherwise shows the displayed structure?  These
>>>> examples work okay:
>>>>
>>>>
>>>> (%i28) block([inflag : true], lisptree(a/b));
>>>> (%o28)
>>>> *__.
>>>> |  |
>>>> a  ^__.
>>>>    |  |
>>>>    b  -1
>>>>
>>>> (%i29) block([inflag : false], lisptree(a/b));
>>>> (%o29)
>>>> /__.
>>>> |  |
>>>> a  b
>>>> (%i30)
>>>>
>>>> _______________________________________________
>> Maxima-discuss mailing list
>> Max...@li...
>> https://lists.sourceforge.net/lists/listinfo/maxima-discuss
>>
>