[REBOL] Re: Is there a REBOL construct similar to the "map" word as used in Logo
From: lmecir:mbox:vol:cz at: 28-Jun-2002 12:23
Hi Gerard,
you may find MAP and more interesting functions in
http://www.rebolforces.com/~ladislav/highfun.r Moreover, you can have a look
at http://www.rebol.cz/cffr.html
-L
----- Original Message -----
From: "Gerard Cote"
Hello,
First I'd like to thank everybody that helped me elucidate my previous
question about the use of the SET word for doing multiple assignments from 2
blocks, specially Joel for the detailed explanation he furnished to me.
Now is another question relative to the word "map" I already used in Logo
for working on agregates - that is data collections - and doesn't seem to
exist per se in the same form under REBOL.
May be I don't really need it for solving the following case but
nevertheless I would like to get some advice about something similar in
REBOL if anyone exists that is different than the two ones I found myself
when using the "forall" and "foreach" words.
Here is an example of what I am just trying to get when using the Logo "map"
construct :
Extract the initials of a complete name by getting the first letter from
each of the first and the last names contained in a block. Using the REBOL
syntax this could be done with the following statement :
>> print initials [George Harrison]
[G H]
Under Logo, the "map" statement is used to work on complete data collections
by using a single statement that shows a very concise syntax - from which
stems my interest for this "map" word.
For example to extract the initials above I can put "map" to work in a
function named "initials" as this :
to initials :name <-- Here "to" plays the same
role as the "func" word under REBOL
output map "first :name <-- "output" simply ends the function
and returns the resulting value
end
to the current context
A detailed explanation
of how the function operates appears below.
After the "initials" function has been defined, the user of the program can
call it with the command that follows after the ? Logo prompt below :
? show initials [George Harrison] "show" is similar to the
print
in REBOL.
[G H] <--- Results
returned when "initials" is evaluated.
This simply defines a function named "initials" that accepts an input
parameter called "name"
the body itself is trying to display something that the "map" statement
applies to each member of the input parameter "name". Here this "something"
is the action represented by the lit-word "first".
It actually takes the first letter of each word contained in the list
submitted as a parameter to the initials function and return all of them in
a block enclosed in square brackets.
I know that I can use "foreach" and "forall" to solve this problem (look
below for my solutions) but can somebody tell me if there is a REBOL word
able to follow very closely the syntax used by the Logo "map" construct ?
I already tried the following alternatives that all seem to work as expected
for my previous problem :
>>my-sentence: [this is a simple sentence]
== [this is a simple sentence]
>> foreach word my-sentence [print first to-string word]
t
i
a
s
s
>> forall my-sentence [print first to-string my-sentence]
t
i
a
s
s
The sole reason I plan to look further in the direction of a REBOL "map"
clone is that I want to
experiment some other concept that looks very promising and seems more
powerful than what I actually know. The man who is the lead developer for
the UCB-Logo I am also playing with while learning REBOL - Mr. Brian
Harvey - names this promising and powerful concept a "template-based
iteration".
Other functional languages also frequently use the term "anonymous function"
for something that looks a bit related to the question-marked items in the
following examples but it seems that's all there is to it, as far as I know
of.
In fact I already found under REBOL some ways to achieve results similar to
those offered by the Logo "map" construct. But the actual syntax I use is
longer and has a different form than the one I look for (maybe I am only
dreaming awake... with the hope to get something that don't really exists
under REBOL !!!).
Whatever the case, efficiently achieving the same results would probably
require to use the PARSE word but I am just curious to know if someone is
aware about the existence of any current REBOL construct for achieving
similar results for the simple cases provided below.
This REBOL construct - if it really exists - would completely eliminate the
need for me to write even the simplest parsing rules for such trivial cases.
Here are 2 specific obvious uses of the Logo "map" construct as I would like
it to be in REBOL too :
? show map [? * ?] [2 3 4 5] <-- Applies the
product of an item with itself to each member of
[4 9 16 25] the
second list.
? show (map [word ?1 ?2 ?1] [a b c] [d e f]) <--- this means create words
that are combinations of 3 letters
[ada beb cfc] the
first and third ones being taken from the first list while
the second onecomes from the second list.
Thank you in advance for any information relative to such a construct if
any.
Regards,
Gerard
P.S. If nobody has heard of a REBOL clone for this "map" construct, I plan
to write one myself since I want to experiment with these concepts under
REBOL exactly as it is possible to do so with Logo. If I am succesful with
the implementation and I find this is really unavoidable to work without it,
I'll share it with everybody interested by this.
