Smalltalk › Squeak › Squeak - Dev

The Object>>recurse: Method

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

5 messages Options

Kjell Godo

The Object>>recurse: Method

Object>>recurse: aZeroInputValuable"ThatReturnsATwoInputValuable"

^( aZeroInputValuable value

) value:( self )value:[ :anObject | anObject recurse: aZeroInputValuable ]

"Object>>recurse:

does for recursion what >>do: does for iteration

it enables any Object O to do recursion

Without writing a new recursive method in the Class of O

It does this even if the

BlockClosures in this version of Smalltalk being used

Can only do tail recursion accurately

Like in Dolphin Smalltalk 6

You call it like

( anObject recurse:[[ :anobject :recurse |

( anobject someChildren collect:[ :child |

( something something ).

( ( recurse value:( child something ) ) something )

]

) asSomeObjectSomething

]] )

So i haven't actually tried it yet

so i do not actually know if it actually works

so it's too hot off the presses for that

because I'm more into a math major than physics

i guess i always like math way better

and so i have a tendency to say that has already been shown

but i can't see how it can fail

So i am putting it out there right now without testing it first

each recursion creates a new twoInputValuable

and the >>recurse: method keeps getting called

from inside of itself

so i think it works

but i guess you have to try it to find out

if you can figure an easier way

or why it can't work

then please show it

i have been trying to do this for a long time

and it shocks me that it's so small must be something wrong in there

so if you easily find a better way

or that it's no way at all

then i missed it and I'm too early

and I'm semi mal informed

ie stupeid

and I'm a baaad scientrist cepting I'm not one

et all

not quite easily done

Kjell E Godo"

Frank Shearar-3

Re: The Object>>recurse: Method

On 27 May 2016 at 10:42, Kjell Godo <[hidden email]> wrote:

Object>>recurse: aZeroInputValuable"ThatReturnsATwoInputValuable"

^( aZeroInputValuable value
) value:( self )value:[ :anObject | anObject recurse: aZeroInputValuable ]

"Object>>recurse:

does for recursion what >>do: does for iteration

it enables any Object O to do recursion
Without writing a new recursive method in the Class of O

Yep. Unless my eyes deceive me, that's the Y combinator. You can see a more classically academic implementation at https://rosettacode.org/wiki/Y_combinator#Smalltalk

frank

timrowledge

Re: The Object>>recurse: Method

> On 27-05-2016, at 1:55 PM, Frank Shearar <[hidden email]> wrote:
> You can see a more classically academic implementation at https://rosettacode.org/wiki/Y_combinator#Smalltalk

Good grief. I didn’t know there are so many ugly languages that I hope never to see again...
tim
--
tim Rowledge; [hidden email]; http://www.rowledge.org/tim
Fractured Idiom:- J'Y SUIS, J'Y PESTES - I can stay for the weekend

marcel.taeumel

Re: The Object>>recurse: Method

In reply to this post by Kjell Godo

Kjell Godo wrote

Object>>recurse: aZeroInputValuable"ThatReturnsATwoInputValuable"

^( aZeroInputValuable value
) value:( self )value:[ :anObject | anObject recurse: aZeroInputValuable ]

"Object>>recurse:

does for recursion what >>do: does for iteration

it enables any Object O to do recursion
Without writing a new recursive method in the Class of O

It does this even if the
BlockClosures in this version of Smalltalk being used
Can only do tail recursion accurately
Like in Dolphin Smalltalk 6

You call it like

( anObject recurse:[[ :anobject :recurse |
( anobject someChildren collect:[ :child |
( something something ).
( ( recurse value:( child something ) ) something )
]
) asSomeObjectSomething
]] )

ok
So i haven't actually tried it yet
so i do not actually know if it actually works
so it's too hot off the presses for that
because I'm more into a math major than physics
i guess i always like math way better
and so i have a tendency to say that has already been shown
but i can't see how it can fail
So i am putting it out there right now without testing it first
each recursion creates a new twoInputValuable
and the >>recurse: method keeps getting called
from inside of itself

so i think it works

but i guess you have to try it to find out

if you can figure an easier way
or why it can't work
then please show it

i have been trying to do this for a long time
and it shocks me that it's so small must be something wrong in there
so if you easily find a better way
or that it's no way at all
then i missed it and I'm too early
and I'm semi mal informed
ie stupeid
and I'm a baaad scientrist cepting I'm not one
et all
not quite easily done

Kjell E Godo"

Hi Kjell,

I suppose that this code is a suggestion for Trunk? If you did not try it out, where does it come from? Dolphin Smalltalk 6?

It does look like a valuable addition to me. Although, I read it like re-curse and thought of Monkey Island and LeChuck. :-D Are there other words for this?

#combine:
#doRecursive:

Anyway, it seems to work for Fibonacci:

6 recurse: [[:number :step |
number caseOf: {
[0] -> [1].
[1] -> [1].
} otherwise: [
(step value: number - 2)
+ (step value: number - 1)]]].

Best,
Marcel

Frank Shearar-3

Re: The Object>>recurse: Method

On 29 May 2016 at 02:04, marcel.taeumel <[hidden email]> wrote:

Kjell Godo wrote

> Object>>recurse: aZeroInputValuable"ThatReturnsATwoInputValuable"
>
> ^( aZeroInputValuable value
> ) value:( self )value:[ :anObject | anObject recurse: aZeroInputValuable
> ]
>
> "Object>>recurse:
>
> does for recursion what >>do: does for iteration
>
> it enables any Object O to do recursion
> Without writing a new recursive method in the Class of O
>
> It does this even if the
> BlockClosures in this version of Smalltalk being used
> Can only do tail recursion accurately
> Like in Dolphin Smalltalk 6
>
> You call it like
>
> ( anObject recurse:[[ :anobject :recurse |
> ( anobject someChildren collect:[ :child |
> ( something something ).
> ( ( recurse value:( child something ) ) something )
> ]
> ) asSomeObjectSomething
> ]] )
>
> ok
> So i haven't actually tried it yet
> so i do not actually know if it actually works
> so it's too hot off the presses for that
> because I'm more into a math major than physics
> i guess i always like math way better
> and so i have a tendency to say that has already been shown
> but i can't see how it can fail
> So i am putting it out there right now without testing it first
> each recursion creates a new twoInputValuable
> and the >>recurse: method keeps getting called
> from inside of itself
>
> so i think it works
>
> but i guess you have to try it to find out
>
> if you can figure an easier way
> or why it can't work
> then please show it
>
> i have been trying to do this for a long time
> and it shocks me that it's so small must be something wrong in there
> so if you easily find a better way
> or that it's no way at all
> then i missed it and I'm too early
> and I'm semi mal informed
> ie stupeid
> and I'm a baaad scientrist cepting I'm not one
> et all
> not quite easily done
>
> Kjell E Godo"

Hi Kjell,

I suppose that this code is a suggestion for Trunk? If you did not try it
out, where does it come from? Dolphin Smalltalk 6?

It does look like a valuable addition to me. Although, I read it like
re-curse and thought of Monkey Island and LeChuck. :-D Are there other words
for this?

#combine:
#doRecursive:

Anyway, it seems to work for Fibonacci:

6 recurse: [[:number :step |
number caseOf: {
[0] -> [1].
[1] -> [1].
} otherwise: [
(step value: number - 2)
+ (step value: number - 1)]]].

I bet you'd like it a whole lot less if someone used it and you had to debug it. I've done things that were at most half as tricky as this [1], and it was NOT fun to debug. In fact, trying to do so breaks the debugger if you do the obvious thing and try debug-edit-and-resume, and I still owe a fix for that...

frank

[1] https://github.com/frankshearar/Parsing-Derivatives/blob/master/Derivatives/DerivingParser.st#L199..L226

Best,
Marcel

--
View this message in context: http://forum.world.st/The-Object-recurse-Method-tp4897901p4897988.html
Sent from the Squeak - Dev mailing list archive at Nabble.com.