[Jprogramming] Fold -- can't get it to work

[Piet de Jong]

My issue is not with x and y insofar as (u F:. v) is concerned.   I agree it  makes sense for x to be 'control information/initial condition' and y the ’data’.

My concern is with x and y insofar as v is concerned.    Each item in y for  the verb (u F:. v) now becomes x for the verb v.    So the data items y become x’s from v’s point of view.  

Thanks for making this great addition to J   (ie the conjunction F:. etc).    

> On 2 Jul 2020, at 10:32 pm, Henry Rich <henryhrich at gmail.com> wrote:
> 
> The J way is for x to be 'control information' and y to be 'data', so that x u^:v y is intuitive.
> 
> Also, u/ processing is generally right to left through a cascade of items.
> 
> Also, x is optional.  y can never be optional.
> 
> Putting those thoughts together, I came up with the current design.  Some agree, some have complained, but not stridently enough to convince us to change it.  Dissenters could get in the habit of writing (v~), perhaps.
> 
> Henry Rich
> 
> On 7/2/2020 3:22 AM, Piet de Jong wrote:
>> Thanks.   Didn’t realise the significance of the change.  Have now got it to work.
>> 
>> There is a feature of F:.  which I find confusing.
>> 
>> The x  argument to  u F:. v appears to be the initial condition while the items of y (I understand) are the items to be processed.
>> 
>> However insofar as  v is concerned, the roles of x and y are reversed.   That is each y  item   becomes x for  v as while the result of the previous application of v is assumed to be  y.   Why this design?
>> 
>> Isn’t it more natural to think of the result of each iteration as the new “initial condition.”   (ie x) and the y for v as the next item to be processed — ie the relevant item of y   as passed to (u F:. v)
>> 
>> Having said that I find F:. an impressive addition.
>>   
>>> On 2 Jul 2020, at 10:14 am, Henry Rich <henryhrich at gmail.com> wrote:
>>> 
>>> Fold changed in 9.02, reversing the meaning of u and v.  You have 9.01 and are expecting the old interpretation.  NuVoc is for 9.02 (it has a note to that effect IIRC).
>>> 
>>> Henry Rich
>>> 
>>> On 7/1/2020 7:29 PM, Piet de Jong wrote:
>>>> Started looking at fold — seems a very useful new verb.
>>>> However couldn’t get it to work
>>>> 
>>>> Engine: j901/j64avx2/darwin
>>>> Release-e: commercial/2020-01-29T12:41:32
>>>> Library: 9.01.21
>>>> Qt IDE: 1.8.6/5.12.7(5.12.7)
>>>> Platform: Darwin 64
>>>> Installer: J901 install
>>>> InstallPath: /applications/j901
>>>> 
>>>> Here is the what happens when I run the example from the vocabulary
>>>> 
>>>>    u
>>>> 3 : 0
>>>> z=. -y
>>>> z [smoutput ' u' ; y ; '-->' ; z
>>>> )
>>>> v
>>>> 4 : 0
>>>> z=. y + 0.01
>>>> z [smoutput x ; 'v' ; y ; '-->' ; z
>>>> )
>>>> y
>>>> 100 0.1 1.1 2.1 3.1
>>>> u F:.v y
>>>> |domain error
>>>> | u F:.v y
>>>> 
>>>> Any help much appreciated.
>>>> 
>>>> 
>>>> ----------------------------------------------------------------------
>>>> For information about J forums see http://www.jsoftware.com/forums.htm
>>> 
>>> -- 
>>> This email has been checked for viruses by AVG.
>>> https://www.avg.com
>>> 
>>> ----------------------------------------------------------------------
>>> For information about J forums see http://www.jsoftware.com/forums.htm
>> ----------------------------------------------------------------------
>> For information about J forums see http://www.jsoftware.com/forums.htm
> 
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm