[proposal] unification: using '() as list identity deref arg

[hartrock]

[update] -> 10.5.6.



During writing some code to ease navigation in lists, I've stumbled about something. Let's speak code for itself.



Assumed there is a definition for ref-parent and some example list:



;; nil for non-existing parent

Code Select Expand


(define (ref-parent r)
  (if (null? r) nil (0 -1 r)))
;;
(setq l_nested '("1" "2" "3"
                  ("4" "5" "6"
                   ("7" "8" "9"))
                 "10" "11" "12"
                  ("13" "14" "15"
                   ("16" "17" "18")))
      )

Now it's possible to eval:

Code Select Expand


> (l_nested (ref-parent (ref "14" l_nested)))
("13" "14" "15" ("16" "17" "18"))

; but it is not possible to eval:

Code Select Expand


> (l_nested (ref-parent (ref "11" l_nested)))

ERR: missing argument

Reason is, that:

Code Select Expand


> (ref-parent (ref "14" l_nested))
(7)

; but:

Code Select Expand


> (ref-parent (ref "11" l_nested))
()

: here '() is not being accepted as deref argument.



What about introducing '() as allowed deref argument, which dereferences list itself to which it is applied to?

Then the following would work:

Code Select Expand


> (l_nested (ref-parent (ref "11" l_nested)))
("1" "2" "3" ("4" "5" "6" ("7" "8" "9")) "10" "11" "12" ("13" "14" "15" ("16" "17" "18")))


This seems to be a neat unification regarding dereferencing lists by refs.



Because ref'ing a non-existing element gives:

Code Select Expand


> (ref "not there" l_nested)
nil

; '() seems to be free for this purpose.

If there is:

Code Select Expand


> (ref-all "not there" l_nested)
()

, we get a '(). But this is no problem, because for dereferencing refs gotten this way, we have to iterate through this list, which would do nothing here.



Are there any problems with this proposal, which I don't see?



What do you think?



Feedback to this proposal is appreciated (as usual).

[Lutz]

I like the idea: http://www.newlisp.org/downloads/development/inprogress/CHANGES-10.5.6.txt">http://www.newlisp.org/downloads/develo ... 10.5.6.txt">http://www.newlisp.org/downloads/development/inprogress/CHANGES-10.5.6.txt



... working as expected on your example.

[hartrock]

Thanks: it works like a charm :-)

Could be a speed record from making a proposal to getting it done... :-)

[xytroxon]

No...



He usually adds the functionality before you have fully described (or understood), what functionality it is that you want added...



Then to add insult to injury, his solution works better and is more elegant... (Although it may take you a day or two to fully realize it ;o)



-- xytroxon

[xytroxon]

Case in point, from this weekends inprogress  newlisp-10.5.6.tgz



I was just thinking about how integer numbers could be used as hashkeys, and I find this addition in  CHANGES-10.5.6.txt

Integers are accepted as hash keys. This allows creating sparse vectors:

       (new Tree 'V)

       (V 123 "hello")

       (V 123) => "hello"

( I thought Lutz's PhD was in psychology, not parapsychology!!! ;o)



-- xytroxon

[TedWalther]

Case in point, from this weekends inprogress  newlisp-10.5.6.tgz



I was just thinking about how integer numbers could be used as hashkeys, and I find this addition in  CHANGES-10.5.6.txt

Integers are accepted as hash keys. This allows creating sparse vectors:

       (new Tree 'V)

       (V 123 "hello")

       (V 123) => "hello"

( I thought Lutz's PhD was in psychology, not parapsychology!!! ;o)



-- xytroxon

Sweet!