Re: [Jprogramming] (no subject)

[Erling Hellenäs]

Hi all !

   a=: 2 5 7
   f=: 1 : '/:~(#: }. i. 2 ^ # y) ([: u/ #) /"1 y'
   + f a
2 5 7 7 9 12 14
   * f a
2 5 7 10 14 35 70

Cheers,
Erling

On 2017-10-02 19:45, Erling Hellenäs wrote:

Hi all!

You don't want the combination of zero elements?

   a=: 2 5 7
   /:~(#: i. 2 ^ # a) ([:+/#) /"1 1 a
0 2 5 7 7 9 12 14
   /:~(#: i. 2 ^ # a) ([:*/#) /"1 1 a
1 2 5 7 10 14 35 70

Guess I could drop it.

Cheers,

Erling

On 2017-10-02 18:49, Skip Cave wrote:
Given a set of integers, what is the most concise and or efficient 
way to
list the numbers along with the sum of all combinations of the 
numbers? the

products of all combinations?

for example:

a =. 2 5 7
+ f a    NB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
2 5 7 7 9 12 14

* f a    NB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
2 5 7 10 14 35 70

The function 'f' should work for any verb and any size right argument 
noun

vector.

Skip

Skip Cave
Cave Consulting LLC
--
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

Re: [Jprogramming] (no subject)

[Roger Hui]

http://code.jsoftware.com/wiki/Essays/Divisors

What you have been asking for are computations on all possible subsets.
When dealing with all possible subsets, start with the boolean matrix
#:@i.@(2&^)@#
 .  You can then do B +/ .* x for all sums or B */ . (^~) x for all
products.  Combinations are not the way to go because m comb n will get you
the indices for the subsets of size m, whereas you want subsets of size 0,
1, ..., n.



On Mon, Oct 2, 2017 at 1:03 PM, Skip Cave  wrote:

> My initial goal was to find all the divisors of a number, eg 100:
> Get the prime factors:
> q:100
> 2 2 5 5
>
> Get all combinations of those prime factors (the power set)
>
> ps0 =: #:@i.@(2&^)@#<@#"1 _]
>
> ps0 q:100
>
> ┌┬─┬─┬───┬─┬───┬───┬─┬─┬───┬───┬─┬───┬─┬─┬───┐
>
> ││5│5│5 5│2│2 5│2 5│2 5 5│2│2 5│2 5│2 5 5│2 2│2 2 5│2 2 5│2 2 5 5│
>
> └┴─┴─┴───┴─┴───┴───┴─┴─┴───┴───┴─┴───┴─┴─┴───┘
>
> Get the products of each box:
>
>   */ each ps0 q: 100
>
> ┌─┬─┬─┬──┬─┬──┬──┬──┬─┬──┬──┬──┬─┬──┬──┬───┐
>
> │1│5│5│25│2│10│10│50│2│10│10│50│4│20│20│100│
>
> └─┴─┴─┴──┴─┴──┴──┴──┴─┴──┴──┴──┴─┴──┴──┴───┘
>
>
> Get the unique set of divisors and sort them:
>
>   /:~ ~.>*/ each ps0 q:100
>
> 1 2 4 5 10 20 25 50 100
>
>
> Now thanks to everyone's help, I can find all the divisors of any number:
>
>   f=.3 :'/:~ ~.>*/ each ps0 q:y'
>
>   f 110
>
> 1 2 5 10 11 22 55 110
>
>   f 43
>
> 1 43
>
>   f 444
>
> 1 2 3 4 6 12 37 74 111 148 222 444
>
>
> Skip
>
>
>
> Skip Cave
> Cave Consulting LLC
>
> On Mon, Oct 2, 2017 at 2:40 PM, Raul Miller  wrote:
>
> > Perhaps this is closer to what you want?
> >
> > F=:1 :', u@> { (~. (u@#"0~ i.@>:)&.> #/.~) y'
> > f=: /F
> >
> > *f 2 2 5 7
> > 1 7 5 35 2 14 10 70 4 28 20 140
> > +f 2 2 5 7
> > 0 7 5 12 2 9 7 14 4 11 9 16
> >
> > I probably should just merge the two definitions together, but I think
> > this works?
> >
> > (Note that I have sort of assumed that the verb argument to f is an
> > associative verb. If you're doing something where that is not the
> > case... well... I guess I would need more specification.)
> >
> > Thanks,
> >
> > --
> > Raul
> >
> >
> > On Mon, Oct 2, 2017 at 3:17 PM, Skip Cave 
> wrote:
> > > Sorry about the wrong terminology. What i meant was:
> > >
> > > Given a vector of random integers (there may be duplicates), what is
> the
> > > most concise and or efficient way to
> > > generate a list of the numbers along with the sum of all combinations
> of
> > > the numbers in a single vector? How about the products of all
> > > combinations?
> > >
> > > Skip Cave
> > > Cave Consulting LLC
> > >
> > > On Mon, Oct 2, 2017 at 1:51 PM, Raul Miller 
> > wrote:
> > >
> > >> Prime factors of an integer are not, in the general case, a set. And
> > >> you really should be careful to avoid specifying a set when what you
> > >> want is not a set.
> > >>
> > >> You might be interested in
> > >> https://rosettacode.org/wiki/Factors_of_an_integer#J ?
> > >>
> > >> That said, refinement is an important part of the specification
> > >> process, so - since it seems you were looking for something different
> > >> - maybe it's worth redoing the specification?
> > >>
> > >> Thanks,
> > >>
> > >> --
> > >> Raul
> > >>
> > >>
> > >> On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave 
> > wrote:
> > >> > My original approach was even more naive than Marc's:
> > >> >
> > >> > NB. from   Roger Hui's Combinations essay on the J website:
> > >> > https://goo.gl/WL4nXn
> > >> >
> > >> > c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^) NB. Roger called this
> > >> 'comb2'
> > >> >
> > >> > NB. I used this definition because I could cut & paste one line, and
> > not
> > >> > require an editor
> > >> >
> > >> > a =. 2 2 5 5
> > >> >
> > >> > ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
> > >> >
> > >> > 100 20 50 4 10 25 2 5
> > >> >
> > >> >
> > >> > I like to sort it:
> > >> >
> > >> > /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1
> c
> > 4){a
> > >> >
> > >> > 2 4 5 10 20 25 50 100
> > >> >
> > >> >
> > >> > The final goal of this exercise was to find all the divisors of an
> > >> integer
> > >> > (not just the prime divisors).
> > >> >
> > >> >
> > >> > So you need to find the prime factors of the integer, for example
> 100:
> > >> >
> > >> >
> > >> > ]a =. q:100
> > >> >
> > >> > 2 2 5 5
> > >> >
> > >> > /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),,
> > |:(1 c
> > >> 4){a
> > >> >
> > >> > 2 4 5 10 20 25 50 100
> > >> >
> > >> >
> > >> > So these are all the divisors of 100.
> > >> >
> > >> >
> > >> > To use Raul's verb, it needs some mods. I can't do the unique until
> > the
> > >> > end, because prime factors of an integer are often duplicated.
> > >> >
> > >> >
> > >> > F1=:1 :'u@#~ #:@i.@(2^#)'   NB. Here's Raul's 

Re: [Jprogramming] zmq - big step for J socket services

['Pascal Jasmin' via Programming]

There is a different J code path with windows and linux, not to mention 
underlying system differences.

I made a repeat run with (J session that had been dormant for a day and) my 
echo modification, and it (qrun) hung, on the last "end task", and presumably 
on the kill__c command.  I've tried the various other zmq dlls, including debug 
releases, and they all crash eventually on qrun.

An interesting feature about the crashes, the redefined echo has the same 
behaviour as in jconsole (the latter doesn't need "doevents" to print 
immediately) -- is that all of the crashes occur on the last "end task".




From: bill lam 
To: 'Pascal Jasmin' via Programming  
Sent: Monday, October 2, 2017 12:11 AM
Subject: Re: [Jprogramming] zmq - big step for J socket services


jconsole and jhs can not run wd commands and I
don't think problem is qt related or can be solved
by qt.

I had tried commenting the line in jcs_lab.ijs
'must be < 20' assert y<20

also change 'e7' to 'e5'

Then this sentence ran fine on linux64 jconsole 
intel core i-7 quad-core (hyperthread disabled)

(qrun bind 99 99)^:10 ''

Пт, 29 сен 2017, jprogramming написал(а):
> 
> A solid accidental fix comes from changing echo:
> 
> 
> echo_z_  =: (0 0 $1!:2&2)`(wd bind 'msgs' ] 0 0 $ 1!:2&2)@.IFQT
> 
> the liberal amount of wd 'msgs' in qrun, seems to let the system (perhaps 
> down to zmq) catch up to the state it thinks it should be in.
> 
> From: 'Pascal Jasmin' via Programming 
> To: "programm...@jsoftware.com"  
> Sent: Friday, September 29, 2017 3:50 PM
> Subject: Re: [Jprogramming] zmq - big step for J socket services
> 
> 
> 
> 
> 
> changing the kill verb to:
> 
> kill=: 3 : 0 
> access=: su 
> runa'exit 0' 
> killp PORT NB. new line.
> destroy'' 
> i.0 0 
> ) 
> 
> 
> helps with the sockets command consistently returning empty, and in general 
> lengthens the test lifetime but 
> 
> qrun 11 11 will still fail within 3 "reruns"
> 
> (qrun modified to use e6 instead of e7 for convenience)
> 
> windows 10
> 
> 
> 
> From: Eric Iverson 
> To: Programming forum  
> Sent: Friday, September 29, 2017 9:59 AM
> Subject: Re: [Jprogramming] zmq - big step for J socket services
> 
> 
> 
> When you get a server error, lse__c  gives you the J error message. I think
> this is covered in the lab and in the help.
> 
> kill and related things depend on host services and that code is still a
> bit rough and naive. It will be cleaned bup with experience or more
> rigorous study. Problems in this area are do not indicate zmq problems per
> se.
> 
> Tracking down qrun hang would be worthwhile. Again, I will bet it has to do
> with starting/tracking/killing tasks and not directly a zmq problem.
> 
> On Thu, Sep 28, 2017 at 7:14 PM, 'Pascal Jasmin' via Programming <
> programm...@jsoftware.com> wrote:
> 
> > its working pretty well,
> >
> > A suggestion for  generic "server error" would be to return the J error
> > from the J server instead.  Useful for the examples you are showing now.
> > (basically show lse__ as client response)
> >
> >
> >
> > The labs and examples do eventually hang if repeated enough though.
> > A repeated crash on the zmq lab occurs after the lines
> >
> > kill__c ''
> > kill__d ''
> >
> > servers_jcs_ '' still shows the c port open 65201
> >
> > then "jcst port" hangs.
> >
> > killp_jcs_ 65201 removes it from servers
> > the kill__c command later in the lab does not crash.
> >
> >
> > if qrun is repeated enough times, it will also hang the j session.
> > 
> > From: Eric Iverson 
> > To: Programming forum 
> > Sent: Thursday, September 28, 2017 1:04 PM
> > Subject: [Jprogramming] zmq - big step for J socket services
> >
> >
> >
> > New addons net/zmq and net/jcs are available. They require 806.
> >
> >
> > The zmq addon has bindings between J and the zmq shared library.
> >
> >
> > The jcs addon has J client/server tools built on the zmq request/reply
> >
> > pattern.
> >
> >
> > These addons make it much easier and with far less code to provide
> >
> > high-performance, robust, and scalable socket services.
> >
> >
> > Start with 806-beta-6 installed from the latest zip packages. Then be sure
> >
> > you have the
> >
> > latest base library and the net/zmq and net/jcs addons installed.
> >
> >
> > An easy way to get started is to run the jcs lab.
> >
> >
> > start J
> >
> >load'labs/labs'
> >
> >lab'~addons/net/jcs/jcs_lab.ijs'
> >
> > step through the lab
> >
> >
> > The lab will let you know if you need to install zmq and where to look for
> >
> > info.
> >
> >
> > Sharing zmq install hints in the forum could save others some trouble.
> >
> >
> > ***
> >
> > Among other things, jcs makes it easy to 

Re: [Jprogramming] (no subject)

['Bo Jacoby' via Programming]

f=.[:+/]*[:|:(2#~#)#:[:i.2^#
f 2 5 7

0 7 5 12 2 9 7 14
 g=.[:*/]^[:|:(2#~ #)#:[:i.2^#
 g 2 5 7
1 7 5 35 2 14 10 70
NB. I didn't check if this solution has already been offered. I just coded it.
 

Den 21:40 mandag den 2. oktober 2017 skrev Raul Miller 
:
 

 Perhaps this is closer to what you want?

  F=:1 :', u@> { (~. (u@#"0~ i.@>:)&.> #/.~) y'
  f=: /F

  *f 2 2 5 7
1 7 5 35 2 14 10 70 4 28 20 140
  +f 2 2 5 7
0 7 5 12 2 9 7 14 4 11 9 16

I probably should just merge the two definitions together, but I think
this works?

(Note that I have sort of assumed that the verb argument to f is an
associative verb. If you're doing something where that is not the
case... well... I guess I would need more specification.)

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 3:17 PM, Skip Cave  wrote:
> Sorry about the wrong terminology. What i meant was:
>
> Given a vector of random integers (there may be duplicates), what is the
> most concise and or efficient way to
> generate a list of the numbers along with the sum of all combinations of
> the numbers in a single vector? How about the  products of all
> combinations?
>
> Skip Cave
> Cave Consulting LLC
>
> On Mon, Oct 2, 2017 at 1:51 PM, Raul Miller  wrote:
>
>> Prime factors of an integer are not, in the general case, a set. And
>> you really should be careful to avoid specifying a set when what you
>> want is not a set.
>>
>> You might be interested in
>> https://rosettacode.org/wiki/Factors_of_an_integer#J ?
>>
>> That said, refinement is an important part of the specification
>> process, so - since it seems you were looking for something different
>> - maybe it's worth redoing the specification?
>>
>> Thanks,
>>
>> --
>> Raul
>>
>>
>> On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave  wrote:
>> > My original approach was even more naive than Marc's:
>> >
>> > NB. from    Roger Hui's Combinations essay on the J website:
>> > https://goo.gl/WL4nXn
>> >
>> > c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^)      NB. Roger called this
>> 'comb2'
>> >
>> > NB. I used this definition because I could cut & paste one line, and not
>> > require an editor
>> >
>> >      a =. 2 2 5 5
>> >
>> > ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>> >
>> > 100 20 50 4 10 25 2 5
>> >
>> >
>> > I like to sort it:
>> >
>> >      /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>> >
>> > 2 4 5 10 20 25 50 100
>> >
>> >
>> > The final goal of this exercise was to find all the divisors of an
>> integer
>> > (not just the prime divisors).
>> >
>> >
>> > So you need to find the prime factors of the integer, for example 100:
>> >
>> >
>> >      ]a =. q:100
>> >
>> > 2 2 5 5
>> >
>> >          /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c
>> 4){a
>> >
>> > 2 4 5 10 20 25 50 100
>> >
>> >
>> > So these are all the divisors of 100.
>> >
>> >
>> > To use Raul's verb, it needs some mods. I can't do the unique until the
>> > end, because prime factors of an integer are often duplicated.
>> >
>> >
>> >      F1=:1 :'u@#~ #:@i.@(2^#)'    NB. Here's Raul's verb minus the
>> initial
>> > unique
>> >
>> > */F1 q:100      NB. we take the product
>> >
>> > 1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100
>> >
>> > ~.*/F1 q:100    NB. Now we take the unique
>> >
>> > 1 5 25 2 10 50 4 20 100
>> >
>> > /:~~.*/F1 q:100      NB. Sort it to make it pretty
>> >
>> > 1 2 4 5 10 20 25 50 100
>> >
>> >
>> > Didn't really need the 1, but Raul likes the empty combination.
>> >
>> >
>> > Now we put it all in one verb:
>> >
>> >
>> >      F2=. /:~~.*/F1 q:
>> >
>> >      F2 100
>> >
>> > |length error: F2
>> >
>> > | F2 100
>> >
>> >      F2=. /:~~.*/F1 q:]
>> >
>> >      F2 100
>> >
>> > |domain error: F2
>> >
>> > | F2 100
>> >
>> >
>> > So this is above my pay grade. I'll have to stick with my inline code:
>> >
>> >
>> >      /:~~.*/F1 q:110
>> >
>> > 1 2 5 10 11 22 55 110
>> >
>> >      /:~~.*/F1 q:43
>> >
>> > 1 43
>> >
>> >      /:~~.*/F1 q:45
>> >
>> > 1 3 5 9 15 45
>> >
>> >      /:~~.*/F1 q:444
>> >
>> > 1 2 3 4 6 12 37 74 111 148 222 444
>> >
>> >
>> > So I can find all the divisors of an integer.
>> >
>> >
>> > Skip
>> >
>> >
>> >
>> >
>> >
>> >
>> > Skip Cave
>> > Cave Consulting LLC
>> >
>> > On Mon, Oct 2, 2017 at 12:15 PM, Marc Simpson  wrote:
>> >
>> >> More naive than Raul's approach, first pass using the 'stats' lib:
>> >>
>> >>      a=.2 5 7
>> >>      require'stats'
>> >>      combv=: ] {~ (comb #@])
>> >>      3 combv a
>> >> 2 5 7
>> >>      2 combv a
>> >> 2 5
>> >> 2 7
>> >> 5 7
>> >>      1 combv a
>> >> 2
>> >> 5
>> >> 7
>> >>      combos=: (1 + i.@#) <@combv"0 1 ]
>> >>      combos a
>> >> ┌─┬───┬─┐
>> >> │2│2 5│2 5 7│
>> >> │5│2 7│        │
>> >> │7│5 7│        │
>> >> └─┴───┴─┘
>> >>      f=: 1 : ';u/"1 each combos y'
>> >>      +f a
>> >> 2 5 7 7 9 12 14
>> >>      *f a
>> >> 2 5 7 10 14 35 70
>> >>
>> >> /M
>> >>
>> >> On Mon, Oct 2, 2017 

Re: [Jprogramming] (no subject)

[Skip Cave]

My initial goal was to find all the divisors of a number, eg 100:
Get the prime factors:
q:100
2 2 5 5

Get all combinations of those prime factors (the power set)

ps0 =: #:@i.@(2&^)@#<@#"1 _]

ps0 q:100

┌┬─┬─┬───┬─┬───┬───┬─┬─┬───┬───┬─┬───┬─┬─┬───┐

││5│5│5 5│2│2 5│2 5│2 5 5│2│2 5│2 5│2 5 5│2 2│2 2 5│2 2 5│2 2 5 5│

└┴─┴─┴───┴─┴───┴───┴─┴─┴───┴───┴─┴───┴─┴─┴───┘

Get the products of each box:

  */ each ps0 q: 100

┌─┬─┬─┬──┬─┬──┬──┬──┬─┬──┬──┬──┬─┬──┬──┬───┐

│1│5│5│25│2│10│10│50│2│10│10│50│4│20│20│100│

└─┴─┴─┴──┴─┴──┴──┴──┴─┴──┴──┴──┴─┴──┴──┴───┘


Get the unique set of divisors and sort them:

  /:~ ~.>*/ each ps0 q:100

1 2 4 5 10 20 25 50 100


Now thanks to everyone's help, I can find all the divisors of any number:

  f=.3 :'/:~ ~.>*/ each ps0 q:y'

  f 110

1 2 5 10 11 22 55 110

  f 43

1 43

  f 444

1 2 3 4 6 12 37 74 111 148 222 444


Skip



Skip Cave
Cave Consulting LLC

On Mon, Oct 2, 2017 at 2:40 PM, Raul Miller  wrote:

> Perhaps this is closer to what you want?
>
> F=:1 :', u@> { (~. (u@#"0~ i.@>:)&.> #/.~) y'
> f=: /F
>
> *f 2 2 5 7
> 1 7 5 35 2 14 10 70 4 28 20 140
> +f 2 2 5 7
> 0 7 5 12 2 9 7 14 4 11 9 16
>
> I probably should just merge the two definitions together, but I think
> this works?
>
> (Note that I have sort of assumed that the verb argument to f is an
> associative verb. If you're doing something where that is not the
> case... well... I guess I would need more specification.)
>
> Thanks,
>
> --
> Raul
>
>
> On Mon, Oct 2, 2017 at 3:17 PM, Skip Cave  wrote:
> > Sorry about the wrong terminology. What i meant was:
> >
> > Given a vector of random integers (there may be duplicates), what is the
> > most concise and or efficient way to
> > generate a list of the numbers along with the sum of all combinations of
> > the numbers in a single vector? How about the products of all
> > combinations?
> >
> > Skip Cave
> > Cave Consulting LLC
> >
> > On Mon, Oct 2, 2017 at 1:51 PM, Raul Miller 
> wrote:
> >
> >> Prime factors of an integer are not, in the general case, a set. And
> >> you really should be careful to avoid specifying a set when what you
> >> want is not a set.
> >>
> >> You might be interested in
> >> https://rosettacode.org/wiki/Factors_of_an_integer#J ?
> >>
> >> That said, refinement is an important part of the specification
> >> process, so - since it seems you were looking for something different
> >> - maybe it's worth redoing the specification?
> >>
> >> Thanks,
> >>
> >> --
> >> Raul
> >>
> >>
> >> On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave 
> wrote:
> >> > My original approach was even more naive than Marc's:
> >> >
> >> > NB. from   Roger Hui's Combinations essay on the J website:
> >> > https://goo.gl/WL4nXn
> >> >
> >> > c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^) NB. Roger called this
> >> 'comb2'
> >> >
> >> > NB. I used this definition because I could cut & paste one line, and
> not
> >> > require an editor
> >> >
> >> > a =. 2 2 5 5
> >> >
> >> > ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
> >> >
> >> > 100 20 50 4 10 25 2 5
> >> >
> >> >
> >> > I like to sort it:
> >> >
> >> > /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c
> 4){a
> >> >
> >> > 2 4 5 10 20 25 50 100
> >> >
> >> >
> >> > The final goal of this exercise was to find all the divisors of an
> >> integer
> >> > (not just the prime divisors).
> >> >
> >> >
> >> > So you need to find the prime factors of the integer, for example 100:
> >> >
> >> >
> >> > ]a =. q:100
> >> >
> >> > 2 2 5 5
> >> >
> >> > /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),,
> |:(1 c
> >> 4){a
> >> >
> >> > 2 4 5 10 20 25 50 100
> >> >
> >> >
> >> > So these are all the divisors of 100.
> >> >
> >> >
> >> > To use Raul's verb, it needs some mods. I can't do the unique until
> the
> >> > end, because prime factors of an integer are often duplicated.
> >> >
> >> >
> >> > F1=:1 :'u@#~ #:@i.@(2^#)'   NB. Here's Raul's verb minus
> the
> >> initial
> >> > unique
> >> >
> >> > */F1 q:100 NB. we take the product
> >> >
> >> > 1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100
> >> >
> >> > ~.*/F1 q:100   NB. Now we take the unique
> >> >
> >> > 1 5 25 2 10 50 4 20 100
> >> >
> >> > /:~~.*/F1 q:100 NB. Sort it to make it pretty
> >> >
> >> > 1 2 4 5 10 20 25 50 100
> >> >
> >> >
> >> > Didn't really need the 1, but Raul likes the empty combination.
> >> >
> >> >
> >> > Now we put it all in one verb:
> >> >
> >> >
> >> > F2=. /:~~.*/F1 q:
> >> >
> >> > F2 100
> >> >
> >> > |length error: F2
> >> >
> >> > | F2 100
> >> >
> >> > F2=. /:~~.*/F1 q:]
> >> >
> >> > F2 100
> >> >
> >> > |domain error: F2
> >> >
> >> > | F2 100
> >> >
> >> >
> >> > So this is above my pay grade. I'll have to stick with my 

Re: [Jprogramming] (no subject)

[Raul Miller]

Perhaps this is closer to what you want?

   F=:1 :', u@> { (~. (u@#"0~ i.@>:)&.> #/.~) y'
   f=: /F

   *f 2 2 5 7
1 7 5 35 2 14 10 70 4 28 20 140
   +f 2 2 5 7
0 7 5 12 2 9 7 14 4 11 9 16

I probably should just merge the two definitions together, but I think
this works?

(Note that I have sort of assumed that the verb argument to f is an
associative verb. If you're doing something where that is not the
case... well... I guess I would need more specification.)

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 3:17 PM, Skip Cave  wrote:
> Sorry about the wrong terminology. What i meant was:
>
> Given a vector of random integers (there may be duplicates), what is the
> most concise and or efficient way to
> generate a list of the numbers along with the sum of all combinations of
> the numbers in a single vector? How about the   products of all
> combinations?
>
> Skip Cave
> Cave Consulting LLC
>
> On Mon, Oct 2, 2017 at 1:51 PM, Raul Miller  wrote:
>
>> Prime factors of an integer are not, in the general case, a set. And
>> you really should be careful to avoid specifying a set when what you
>> want is not a set.
>>
>> You might be interested in
>> https://rosettacode.org/wiki/Factors_of_an_integer#J ?
>>
>> That said, refinement is an important part of the specification
>> process, so - since it seems you were looking for something different
>> - maybe it's worth redoing the specification?
>>
>> Thanks,
>>
>> --
>> Raul
>>
>>
>> On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave  wrote:
>> > My original approach was even more naive than Marc's:
>> >
>> > NB. from Roger Hui's Combinations essay on the J website:
>> > https://goo.gl/WL4nXn
>> >
>> > c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^)   NB. Roger called this
>> 'comb2'
>> >
>> > NB. I used this definition because I could cut & paste one line, and not
>> > require an editor
>> >
>> >   a =. 2 2 5 5
>> >
>> > ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>> >
>> > 100 20 50 4 10 25 2 5
>> >
>> >
>> > I like to sort it:
>> >
>> >   /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>> >
>> > 2 4 5 10 20 25 50 100
>> >
>> >
>> > The final goal of this exercise was to find all the divisors of an
>> integer
>> > (not just the prime divisors).
>> >
>> >
>> > So you need to find the prime factors of the integer, for example 100:
>> >
>> >
>> >   ]a =. q:100
>> >
>> > 2 2 5 5
>> >
>> >   /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c
>> 4){a
>> >
>> > 2 4 5 10 20 25 50 100
>> >
>> >
>> > So these are all the divisors of 100.
>> >
>> >
>> > To use Raul's verb, it needs some mods. I can't do the unique until the
>> > end, because prime factors of an integer are often duplicated.
>> >
>> >
>> >   F1=:1 :'u@#~ #:@i.@(2^#)' NB. Here's Raul's verb minus the
>> initial
>> > unique
>> >
>> > */F1 q:100   NB. we take the product
>> >
>> > 1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100
>> >
>> > ~.*/F1 q:100 NB. Now we take the unique
>> >
>> > 1 5 25 2 10 50 4 20 100
>> >
>> > /:~~.*/F1 q:100   NB. Sort it to make it pretty
>> >
>> > 1 2 4 5 10 20 25 50 100
>> >
>> >
>> > Didn't really need the 1, but Raul likes the empty combination.
>> >
>> >
>> > Now we put it all in one verb:
>> >
>> >
>> >   F2=. /:~~.*/F1 q:
>> >
>> >   F2 100
>> >
>> > |length error: F2
>> >
>> > | F2 100
>> >
>> >   F2=. /:~~.*/F1 q:]
>> >
>> >   F2 100
>> >
>> > |domain error: F2
>> >
>> > | F2 100
>> >
>> >
>> > So this is above my pay grade. I'll have to stick with my inline code:
>> >
>> >
>> >   /:~~.*/F1 q:110
>> >
>> > 1 2 5 10 11 22 55 110
>> >
>> >   /:~~.*/F1 q:43
>> >
>> > 1 43
>> >
>> >   /:~~.*/F1 q:45
>> >
>> > 1 3 5 9 15 45
>> >
>> >   /:~~.*/F1 q:444
>> >
>> > 1 2 3 4 6 12 37 74 111 148 222 444
>> >
>> >
>> > So I can find all the divisors of an integer.
>> >
>> >
>> > Skip
>> >
>> >
>> >
>> >
>> >
>> >
>> > Skip Cave
>> > Cave Consulting LLC
>> >
>> > On Mon, Oct 2, 2017 at 12:15 PM, Marc Simpson  wrote:
>> >
>> >> More naive than Raul's approach, first pass using the 'stats' lib:
>> >>
>> >>   a=.2 5 7
>> >>   require'stats'
>> >>   combv=: ] {~ (comb #@])
>> >>   3 combv a
>> >> 2 5 7
>> >>   2 combv a
>> >> 2 5
>> >> 2 7
>> >> 5 7
>> >>   1 combv a
>> >> 2
>> >> 5
>> >> 7
>> >>   combos=: (1 + i.@#) <@combv"0 1 ]
>> >>   combos a
>> >> ┌─┬───┬─┐
>> >> │2│2 5│2 5 7│
>> >> │5│2 7│ │
>> >> │7│5 7│ │
>> >> └─┴───┴─┘
>> >>   f=: 1 : ';u/"1 each combos y'
>> >>   +f a
>> >> 2 5 7 7 9 12 14
>> >>   *f a
>> >> 2 5 7 10 14 35 70
>> >>
>> >> /M
>> >>
>> >> On Mon, Oct 2, 2017 at 10:06 AM, Raul Miller 
>> >> wrote:
>> >> > For a first effort, I would go with
>> >> >
>> >> >   F=:1 :'(u@#~ #:@i.@(2^#))@~.'
>> >> >   f=: /F
>> >> >
>> >> > Hopefully that makes the issues obvious - the 

Re: [Jprogramming] (no subject)

[Skip Cave]

Sorry about the wrong terminology. What i meant was:

Given a vector of random integers (there may be duplicates), what is the
most concise and or efficient way to
generate a list of the numbers along with the sum of all combinations of
the numbers in a single vector? How about the   products of all
combinations?

Skip Cave
Cave Consulting LLC

On Mon, Oct 2, 2017 at 1:51 PM, Raul Miller  wrote:

> Prime factors of an integer are not, in the general case, a set. And
> you really should be careful to avoid specifying a set when what you
> want is not a set.
>
> You might be interested in
> https://rosettacode.org/wiki/Factors_of_an_integer#J ?
>
> That said, refinement is an important part of the specification
> process, so - since it seems you were looking for something different
> - maybe it's worth redoing the specification?
>
> Thanks,
>
> --
> Raul
>
>
> On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave  wrote:
> > My original approach was even more naive than Marc's:
> >
> > NB. from Roger Hui's Combinations essay on the J website:
> > https://goo.gl/WL4nXn
> >
> > c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^)   NB. Roger called this
> 'comb2'
> >
> > NB. I used this definition because I could cut & paste one line, and not
> > require an editor
> >
> >   a =. 2 2 5 5
> >
> > ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
> >
> > 100 20 50 4 10 25 2 5
> >
> >
> > I like to sort it:
> >
> >   /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
> >
> > 2 4 5 10 20 25 50 100
> >
> >
> > The final goal of this exercise was to find all the divisors of an
> integer
> > (not just the prime divisors).
> >
> >
> > So you need to find the prime factors of the integer, for example 100:
> >
> >
> >   ]a =. q:100
> >
> > 2 2 5 5
> >
> >   /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c
> 4){a
> >
> > 2 4 5 10 20 25 50 100
> >
> >
> > So these are all the divisors of 100.
> >
> >
> > To use Raul's verb, it needs some mods. I can't do the unique until the
> > end, because prime factors of an integer are often duplicated.
> >
> >
> >   F1=:1 :'u@#~ #:@i.@(2^#)' NB. Here's Raul's verb minus the
> initial
> > unique
> >
> > */F1 q:100   NB. we take the product
> >
> > 1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100
> >
> > ~.*/F1 q:100 NB. Now we take the unique
> >
> > 1 5 25 2 10 50 4 20 100
> >
> > /:~~.*/F1 q:100   NB. Sort it to make it pretty
> >
> > 1 2 4 5 10 20 25 50 100
> >
> >
> > Didn't really need the 1, but Raul likes the empty combination.
> >
> >
> > Now we put it all in one verb:
> >
> >
> >   F2=. /:~~.*/F1 q:
> >
> >   F2 100
> >
> > |length error: F2
> >
> > | F2 100
> >
> >   F2=. /:~~.*/F1 q:]
> >
> >   F2 100
> >
> > |domain error: F2
> >
> > | F2 100
> >
> >
> > So this is above my pay grade. I'll have to stick with my inline code:
> >
> >
> >   /:~~.*/F1 q:110
> >
> > 1 2 5 10 11 22 55 110
> >
> >   /:~~.*/F1 q:43
> >
> > 1 43
> >
> >   /:~~.*/F1 q:45
> >
> > 1 3 5 9 15 45
> >
> >   /:~~.*/F1 q:444
> >
> > 1 2 3 4 6 12 37 74 111 148 222 444
> >
> >
> > So I can find all the divisors of an integer.
> >
> >
> > Skip
> >
> >
> >
> >
> >
> >
> > Skip Cave
> > Cave Consulting LLC
> >
> > On Mon, Oct 2, 2017 at 12:15 PM, Marc Simpson  wrote:
> >
> >> More naive than Raul's approach, first pass using the 'stats' lib:
> >>
> >>   a=.2 5 7
> >>   require'stats'
> >>   combv=: ] {~ (comb #@])
> >>   3 combv a
> >> 2 5 7
> >>   2 combv a
> >> 2 5
> >> 2 7
> >> 5 7
> >>   1 combv a
> >> 2
> >> 5
> >> 7
> >>   combos=: (1 + i.@#) <@combv"0 1 ]
> >>   combos a
> >> ┌─┬───┬─┐
> >> │2│2 5│2 5 7│
> >> │5│2 7│ │
> >> │7│5 7│ │
> >> └─┴───┴─┘
> >>   f=: 1 : ';u/"1 each combos y'
> >>   +f a
> >> 2 5 7 7 9 12 14
> >>   *f a
> >> 2 5 7 10 14 35 70
> >>
> >> /M
> >>
> >> On Mon, Oct 2, 2017 at 10:06 AM, Raul Miller 
> >> wrote:
> >> > For a first effort, I would go with
> >> >
> >> >   F=:1 :'(u@#~ #:@i.@(2^#))@~.'
> >> >   f=: /F
> >> >
> >> > Hopefully that makes the issues obvious - the specification here calls
> >> > for a result which grows exponentially with the size of the argument
> >> > set.
> >> >
> >> > Also:
> >> >
> >> > The ~. might be extra work, but for typical cases the effort of
> >> > ensuring that the argument is a set is trivial compared to the effort
> >> > of constructing the result.
> >> >
> >> > You did not include the empty combination in your example results, but
> >> > given your specification my initial inclination is to treat that as an
> >> > oversight.
> >> >
> >> > I defined F instead of going straight for f because for testing
> >> > purposes I want to be able to do ( >> >
> >> > Thanks,
> >> >
> >> > --
> >> > Raul
> >> >
> >> >
> >> > On Mon, Oct 2, 2017 at 12:49 PM, Skip Cave 
> >> wrote:
> 

Re: [Jprogramming] (no subject)

[Raul Miller]

Prime factors of an integer are not, in the general case, a set. And
you really should be careful to avoid specifying a set when what you
want is not a set.

You might be interested in
https://rosettacode.org/wiki/Factors_of_an_integer#J ?

That said, refinement is an important part of the specification
process, so - since it seems you were looking for something different
- maybe it's worth redoing the specification?

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 2:09 PM, Skip Cave  wrote:
> My original approach was even more naive than Marc's:
>
> NB. fromRoger Hui's Combinations essay on the J website:
> https://goo.gl/WL4nXn
>
> c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^) NB. Roger called this 'comb2'
>
> NB. I used this definition because I could cut & paste one line, and not
> require an editor
>
> a =. 2 2 5 5
>
> ~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>
> 100 20 50 4 10 25 2 5
>
>
> I like to sort it:
>
> /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>
> 2 4 5 10 20 25 50 100
>
>
> The final goal of this exercise was to find all the divisors of an integer
> (not just the prime divisors).
>
>
> So you need to find the prime factors of the integer, for example 100:
>
>
> ]a =. q:100
>
> 2 2 5 5
>
> /:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a
>
> 2 4 5 10 20 25 50 100
>
>
> So these are all the divisors of 100.
>
>
> To use Raul's verb, it needs some mods. I can't do the unique until the
> end, because prime factors of an integer are often duplicated.
>
>
> F1=:1 :'u@#~ #:@i.@(2^#)'   NB. Here's Raul's verb minus the initial
> unique
>
> */F1 q:100 NB. we take the product
>
> 1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100
>
> ~.*/F1 q:100   NB. Now we take the unique
>
> 1 5 25 2 10 50 4 20 100
>
> /:~~.*/F1 q:100 NB. Sort it to make it pretty
>
> 1 2 4 5 10 20 25 50 100
>
>
> Didn't really need the 1, but Raul likes the empty combination.
>
>
> Now we put it all in one verb:
>
>
> F2=. /:~~.*/F1 q:
>
> F2 100
>
> |length error: F2
>
> | F2 100
>
> F2=. /:~~.*/F1 q:]
>
> F2 100
>
> |domain error: F2
>
> | F2 100
>
>
> So this is above my pay grade. I'll have to stick with my inline code:
>
>
> /:~~.*/F1 q:110
>
> 1 2 5 10 11 22 55 110
>
> /:~~.*/F1 q:43
>
> 1 43
>
> /:~~.*/F1 q:45
>
> 1 3 5 9 15 45
>
> /:~~.*/F1 q:444
>
> 1 2 3 4 6 12 37 74 111 148 222 444
>
>
> So I can find all the divisors of an integer.
>
>
> Skip
>
>
>
>
>
>
> Skip Cave
> Cave Consulting LLC
>
> On Mon, Oct 2, 2017 at 12:15 PM, Marc Simpson  wrote:
>
>> More naive than Raul's approach, first pass using the 'stats' lib:
>>
>> a=.2 5 7
>> require'stats'
>> combv=: ] {~ (comb #@])
>> 3 combv a
>> 2 5 7
>> 2 combv a
>> 2 5
>> 2 7
>> 5 7
>> 1 combv a
>> 2
>> 5
>> 7
>> combos=: (1 + i.@#) <@combv"0 1 ]
>> combos a
>> ┌─┬───┬─┐
>> │2│2 5│2 5 7│
>> │5│2 7│   │
>> │7│5 7│   │
>> └─┴───┴─┘
>> f=: 1 : ';u/"1 each combos y'
>> +f a
>> 2 5 7 7 9 12 14
>> *f a
>> 2 5 7 10 14 35 70
>>
>> /M
>>
>> On Mon, Oct 2, 2017 at 10:06 AM, Raul Miller 
>> wrote:
>> > For a first effort, I would go with
>> >
>> > F=:1 :'(u@#~ #:@i.@(2^#))@~.'
>> > f=: /F
>> >
>> > Hopefully that makes the issues obvious - the specification here calls
>> > for a result which grows exponentially with the size of the argument
>> > set.
>> >
>> > Also:
>> >
>> > The ~. might be extra work, but for typical cases the effort of
>> > ensuring that the argument is a set is trivial compared to the effort
>> > of constructing the result.
>> >
>> > You did not include the empty combination in your example results, but
>> > given your specification my initial inclination is to treat that as an
>> > oversight.
>> >
>> > I defined F instead of going straight for f because for testing
>> > purposes I want to be able to do (> >
>> > Thanks,
>> >
>> > --
>> > Raul
>> >
>> >
>> > On Mon, Oct 2, 2017 at 12:49 PM, Skip Cave 
>> wrote:
>> >> Given a set of integers, what is the most concise and or efficient way
>> to
>> >> list the numbers along with the sum of all combinations of the numbers?
>> the
>> >> products of all combinations?
>> >>
>> >> for example:
>> >>
>> >> a =. 2 5 7
>> >> + f a NB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
>> >> 2 5 7 7 9 12 14
>> >>
>> >> * f a NB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
>> >> 2 5 7 10 14 35 70
>> >>
>> >> The function 'f' should work for any verb and any size right argument
>> noun
>> >> vector.
>> >>
>> >> Skip
>> >>
>> >> Skip Cave
>> >> Cave Consulting LLC
>> >> --
>> >> For information about J forums see http://www.jsoftware.com/forums.htm
>> > --
>> > For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] zmq - big step for J socket services

[Devon McCormick]

I believe I once got wd to work in jconsole by building my own
event-handler but don't think I have a working example to share.

On Mon, Oct 2, 2017 at 12:11 AM, bill lam  wrote:

> jconsole and jhs can not run wd commands and I
> don't think problem is qt related or can be solved
> by qt.
>
> I had tried commenting the line in jcs_lab.ijs
> 'must be < 20' assert y<20
>
> also change 'e7' to 'e5'
>
> Then this sentence ran fine on linux64 jconsole
> intel core i-7 quad-core (hyperthread disabled)
>
>  (qrun bind 99 99)^:10 ''
>
> Пт, 29 сен 2017, jprogramming написал(а):
> >
> > A solid accidental fix comes from changing echo:
> >
> >
> > echo_z_  =: (0 0 $1!:2&2)`(wd bind 'msgs' ] 0 0 $ 1!:2&2)@.IFQT
> >
> > the liberal amount of wd 'msgs' in qrun, seems to let the system
> (perhaps down to zmq) catch up to the state it thinks it should be in.
> > 
> > From: 'Pascal Jasmin' via Programming 
> > To: "programm...@jsoftware.com" 
> > Sent: Friday, September 29, 2017 3:50 PM
> > Subject: Re: [Jprogramming] zmq - big step for J socket services
> >
> >
> >
> >
> >
> > changing the kill verb to:
> >
> > kill=: 3 : 0
> > access=: su
> > runa'exit 0'
> > killp PORT NB. new line.
> > destroy''
> > i.0 0
> > )
> >
> >
> > helps with the sockets command consistently returning empty, and in
> general lengthens the test lifetime but
> >
> > qrun 11 11 will still fail within 3 "reruns"
> >
> > (qrun modified to use e6 instead of e7 for convenience)
> >
> > windows 10
> >
> >
> > 
> > From: Eric Iverson 
> > To: Programming forum 
> > Sent: Friday, September 29, 2017 9:59 AM
> > Subject: Re: [Jprogramming] zmq - big step for J socket services
> >
> >
> >
> > When you get a server error, lse__c  gives you the J error message. I
> think
> > this is covered in the lab and in the help.
> >
> > kill and related things depend on host services and that code is still a
> > bit rough and naive. It will be cleaned bup with experience or more
> > rigorous study. Problems in this area are do not indicate zmq problems
> per
> > se.
> >
> > Tracking down qrun hang would be worthwhile. Again, I will bet it has to
> do
> > with starting/tracking/killing tasks and not directly a zmq problem.
> >
> > On Thu, Sep 28, 2017 at 7:14 PM, 'Pascal Jasmin' via Programming <
> > programm...@jsoftware.com> wrote:
> >
> > > its working pretty well,
> > >
> > > A suggestion for  generic "server error" would be to return the J error
> > > from the J server instead.  Useful for the examples you are showing
> now.
> > > (basically show lse__ as client response)
> > >
> > >
> > >
> > > The labs and examples do eventually hang if repeated enough though.
> > > A repeated crash on the zmq lab occurs after the lines
> > >
> > > kill__c ''
> > > kill__d ''
> > >
> > > servers_jcs_ '' still shows the c port open 65201
> > >
> > > then "jcst port" hangs.
> > >
> > > killp_jcs_ 65201 removes it from servers
> > > the kill__c command later in the lab does not crash.
> > >
> > >
> > > if qrun is repeated enough times, it will also hang the j session.
> > > 
> > > From: Eric Iverson 
> > > To: Programming forum 
> > > Sent: Thursday, September 28, 2017 1:04 PM
> > > Subject: [Jprogramming] zmq - big step for J socket services
> > >
> > >
> > >
> > > New addons net/zmq and net/jcs are available. They require 806.
> > >
> > >
> > > The zmq addon has bindings between J and the zmq shared library.
> > >
> > >
> > > The jcs addon has J client/server tools built on the zmq request/reply
> > >
> > > pattern.
> > >
> > >
> > > These addons make it much easier and with far less code to provide
> > >
> > > high-performance, robust, and scalable socket services.
> > >
> > >
> > > Start with 806-beta-6 installed from the latest zip packages. Then be
> sure
> > >
> > > you have the
> > >
> > > latest base library and the net/zmq and net/jcs addons installed.
> > >
> > >
> > > An easy way to get started is to run the jcs lab.
> > >
> > >
> > > start J
> > >
> > >load'labs/labs'
> > >
> > >lab'~addons/net/jcs/jcs_lab.ijs'
> > >
> > > step through the lab
> > >
> > >
> > > The lab will let you know if you need to install zmq and where to look
> for
> > >
> > > info.
> > >
> > >
> > > Sharing zmq install hints in the forum could save others some trouble.
> > >
> > >
> > > ***
> > >
> > > Among other things, jcs makes it easy to start other J tasks and to use
> > >
> > > them as servers. The lab includes a simple example of starting
> multiple J
> > >
> > > tasks to service a queue of jobs.
> > >
> > >
> > > *** from the jcs lab
> > >
> > > zmq is an api built on top of sockets
> > >
> > >
> > > it is available across platforms and removes almost all the mundane and
> > >
> > > critical chores in 

Re: [Jprogramming] Composition conjunctions

[Marc Simpson]

On Mon, Oct 2, 2017 at 11:13 AM, Erling Hellenäs
 wrote:
> And what would you call these posts. A factual discussion? /Erling

Sure; a discussion with suggestions, differing opinions. I find
dismissing replies as 'spam' a tad rude; let's try and keep
conversation civil and friendly (on all sides)?

Anyway, I've said my piece; don't want to derail this thread further.

Best,
M
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

And what would you call these posts. A factual discussion? /Erling

On 2017-10-02 19:54, Marc Simpson wrote:

On Mon, Oct 2, 2017 at 10:51 AM, Erling Hellenäs
 wrote:

[snip]
Normally, if I get spammed like this, I would assume that what I said WAS
important, and I'd repost it and write more about it.

Referring to on-list replies as 'spam' is rather insulting to those
who take time to respond.
--
For information about J forums see http://www.jsoftware.com/forums.htm



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] (no subject)

[Skip Cave]

My original approach was even more naive than Marc's:

NB. fromRoger Hui's Combinations essay on the J website:
https://goo.gl/WL4nXn

c=: ((= +/"1) |.@:I.@# ]) #:@i.@(2&^) NB. Roger called this 'comb2'

NB. I used this definition because I could cut & paste one line, and not
require an editor

a =. 2 2 5 5

~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a

100 20 50 4 10 25 2 5


I like to sort it:

/:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a

2 4 5 10 20 25 50 100


The final goal of this exercise was to find all the divisors of an integer
(not just the prime divisors).


So you need to find the prime factors of the integer, for example 100:


]a =. q:100

2 2 5 5

/:~~.(*/"1(4 c 4){a),(*/"1(3 c 4){a),(*/"1(2 c 4){a),, |:(1 c 4){a

2 4 5 10 20 25 50 100


So these are all the divisors of 100.


To use Raul's verb, it needs some mods. I can't do the unique until the
end, because prime factors of an integer are often duplicated.


F1=:1 :'u@#~ #:@i.@(2^#)'   NB. Here's Raul's verb minus the initial
unique

*/F1 q:100 NB. we take the product

1 5 5 25 2 10 10 50 2 10 10 50 4 20 20 100

~.*/F1 q:100   NB. Now we take the unique

1 5 25 2 10 50 4 20 100

/:~~.*/F1 q:100 NB. Sort it to make it pretty

1 2 4 5 10 20 25 50 100


Didn't really need the 1, but Raul likes the empty combination.


Now we put it all in one verb:


F2=. /:~~.*/F1 q:

F2 100

|length error: F2

| F2 100

F2=. /:~~.*/F1 q:]

F2 100

|domain error: F2

| F2 100


So this is above my pay grade. I'll have to stick with my inline code:


/:~~.*/F1 q:110

1 2 5 10 11 22 55 110

/:~~.*/F1 q:43

1 43

/:~~.*/F1 q:45

1 3 5 9 15 45

/:~~.*/F1 q:444

1 2 3 4 6 12 37 74 111 148 222 444


So I can find all the divisors of an integer.


Skip






Skip Cave
Cave Consulting LLC

On Mon, Oct 2, 2017 at 12:15 PM, Marc Simpson  wrote:

> More naive than Raul's approach, first pass using the 'stats' lib:
>
> a=.2 5 7
> require'stats'
> combv=: ] {~ (comb #@])
> 3 combv a
> 2 5 7
> 2 combv a
> 2 5
> 2 7
> 5 7
> 1 combv a
> 2
> 5
> 7
> combos=: (1 + i.@#) <@combv"0 1 ]
> combos a
> ┌─┬───┬─┐
> │2│2 5│2 5 7│
> │5│2 7│   │
> │7│5 7│   │
> └─┴───┴─┘
> f=: 1 : ';u/"1 each combos y'
> +f a
> 2 5 7 7 9 12 14
> *f a
> 2 5 7 10 14 35 70
>
> /M
>
> On Mon, Oct 2, 2017 at 10:06 AM, Raul Miller 
> wrote:
> > For a first effort, I would go with
> >
> > F=:1 :'(u@#~ #:@i.@(2^#))@~.'
> > f=: /F
> >
> > Hopefully that makes the issues obvious - the specification here calls
> > for a result which grows exponentially with the size of the argument
> > set.
> >
> > Also:
> >
> > The ~. might be extra work, but for typical cases the effort of
> > ensuring that the argument is a set is trivial compared to the effort
> > of constructing the result.
> >
> > You did not include the empty combination in your example results, but
> > given your specification my initial inclination is to treat that as an
> > oversight.
> >
> > I defined F instead of going straight for f because for testing
> > purposes I want to be able to do ( >
> > Thanks,
> >
> > --
> > Raul
> >
> >
> > On Mon, Oct 2, 2017 at 12:49 PM, Skip Cave 
> wrote:
> >> Given a set of integers, what is the most concise and or efficient way
> to
> >> list the numbers along with the sum of all combinations of the numbers?
> the
> >> products of all combinations?
> >>
> >> for example:
> >>
> >> a =. 2 5 7
> >> + f a NB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
> >> 2 5 7 7 9 12 14
> >>
> >> * f a NB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
> >> 2 5 7 10 14 35 70
> >>
> >> The function 'f' should work for any verb and any size right argument
> noun
> >> vector.
> >>
> >> Skip
> >>
> >> Skip Cave
> >> Cave Consulting LLC
> >> --
> >> 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
>
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Louis de Forcrand]

Beyond wether these tests are useful or not, you should run "ts" with a left 
argument, as a single run which lasts only 0.1 seconds can be somewhat 
imprecise.
ts's left argument n is the number of times to run the string of code, and its 
result will be the total time elapsed divided by n.

Cheers,
Louis

> On 2 Oct 2017, at 19:51, Erling Hellenäs  wrote:
> 
> Hi all!
> 
> If something is not interesting or important, a good general idea could be to 
> not spend forum time on lengthy  discussions about it?
> Normally, if I get spammed like this, I would assume that what I said WAS 
> important, and I'd repost it and write more about it. However, since this is 
> a proprietary mail list I will have to leave that to the  readers.
> 
> Cheers,
> Erling
>> On 2017-10-02 19:26, Don Guinn wrote:
>> Detailed measurements are useful and meaningful if your application is
>> performing badly. But general statements about poor performance in parts of
>> an application that isn't used much is a waste of time.
>> 
>> @ vs. @: is a concern but blanket proclamations is wrong. Today @ performs
>> better than @: most of the time, especially for primitives. But for defined
>> verbs, depending on the design, there may be no performance gain and a lot
>> of wasted memory.
>> 
>> If you have an app with a performance problem then run a tool to find out
>> where you're spending the time. J has such a tool, as do almost all
>> programming languages. Now you know where to spend your time.
>> 
>> And you may find out that the problem is not in J, but in your design.
>> 
>> There is no argument that @ has much room for improvement. Which seems to
>> be what your beef is about. So you're wasting your time and everybody
>> else's time proving the obvious.
>> 
>>> On Oct 2, 2017 10:06 AM, "Erling Hellenäs"  wrote:
>>> 
>>> You are welcome to show us better measurements if you think these
>>> measurements are very important and worthy of a lengthy discussion in the
>>> forum. /Erling
>>> 
 On 2017-10-02 17:53, Raul Miller wrote:
 
 The null case here should be ]
 
 (-@- -: ]) v
 1
 
 Thanks,
 
 
>>> --
>>> 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

--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Marc Simpson]

On Mon, Oct 2, 2017 at 10:51 AM, Erling Hellenäs
 wrote:
> [snip]
> Normally, if I get spammed like this, I would assume that what I said WAS
> important, and I'd repost it and write more about it.

Referring to on-list replies as 'spam' is rather insulting to those
who take time to respond.
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Hi all!

If something is not interesting or important, a good general idea could 
be to not spend forum time on lengthy  discussions about it?
Normally, if I get spammed like this, I would assume that what I said 
WAS important, and I'd repost it and write more about it. However, since 
this is a proprietary mail list I will have to leave that to the  readers.


Cheers,
Erling
On 2017-10-02 19:26, Don Guinn wrote:

Detailed measurements are useful and meaningful if your application is
performing badly. But general statements about poor performance in parts of
an application that isn't used much is a waste of time.

@ vs. @: is a concern but blanket proclamations is wrong. Today @ performs
better than @: most of the time, especially for primitives. But for defined
verbs, depending on the design, there may be no performance gain and a lot
of wasted memory.

If you have an app with a performance problem then run a tool to find out
where you're spending the time. J has such a tool, as do almost all
programming languages. Now you know where to spend your time.

And you may find out that the problem is not in J, but in your design.

There is no argument that @ has much room for improvement. Which seems to
be what your beef is about. So you're wasting your time and everybody
else's time proving the obvious.

On Oct 2, 2017 10:06 AM, "Erling Hellenäs"  wrote:


You are welcome to show us better measurements if you think these
measurements are very important and worthy of a lengthy discussion in the
forum. /Erling

On 2017-10-02 17:53, Raul Miller wrote:


The null case here should be ]

 (-@- -: ]) v
1

Thanks,



--
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

Re: [Jprogramming] (no subject)

[Erling Hellenäs]

Hi all!

You don't want the combination of zero elements?

   a=: 2 5 7
   /:~(#: i. 2 ^ # a) ([:+/#) /"1 1 a
0 2 5 7 7 9 12 14
   /:~(#: i. 2 ^ # a) ([:*/#) /"1 1 a
1 2 5 7 10 14 35 70

Guess I could drop it.

Cheers,

Erling

On 2017-10-02 18:49, Skip Cave wrote:

Given a set of integers, what is the most concise and or efficient way to
list the numbers along with the sum of all combinations of the numbers? the
products of all combinations?

for example:

a =. 2 5 7
+ f aNB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
2 5 7 7 9 12 14

* f aNB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
2 5 7 10 14 35 70

The function 'f' should work for any verb and any size right argument noun
vector.

Skip

Skip Cave
Cave Consulting LLC
--
For information about J forums see http://www.jsoftware.com/forums.htm



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] (no subject)

[Ben Gorte - CITG]

http://code.jsoftware.com/wiki/Essays/Power_Set is an essay by Roger Hui about 
power sets, including a number of ways to generate these, such as:

   ps0 =: #:@i.@(2&^)@# <@#"1 _ ]

   ps0 2 5 7
┌┬─┬─┬───┬─┬───┬───┬─┐
││7│5│5 7│2│2 7│2 5│2 5 7│
└┴─┴─┴───┴─┴───┴───┴─┘

Now your f can be:

   f =: 1 : '}. /:~ ; u/ each ps0 y'
   + f 2 5 7
2 5 7 7 9 12 14
   * f 2 5 7
2 5 7 10 14 35 70

Ben

From: Programming [programming-boun...@forums.jsoftware.com] on behalf of Skip 
Cave [s...@caveconsulting.com]
Sent: Monday, October 02, 2017 18:49
To: programm...@jsoftware.com
Subject: [Jprogramming] (no subject)

Given a set of integers, what is the most concise and or efficient way to
list the numbers along with the sum of all combinations of the numbers? the
products of all combinations?

for example:

a =. 2 5 7
+ f aNB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
2 5 7 7 9 12 14

* f aNB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
2 5 7 10 14 35 70

The function 'f' should work for any verb and any size right argument noun
vector.

Skip

Skip Cave
Cave Consulting LLC
--
For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Raul Miller]

Well... hypothetically speaking, a case might be made for special code
which evaluates -@- with a domain check and then ] -- but I guess we
are not seeing anything useful about that case yet.

And there's a small cost to special code, so we usually don't want it
doing things people should not use.

Thanks,

-- 
Raul

On Mon, Oct 2, 2017 at 1:26 PM, Don Guinn  wrote:
> Detailed measurements are useful and meaningful if your application is
> performing badly. But general statements about poor performance in parts of
> an application that isn't used much is a waste of time.
>
> @ vs. @: is a concern but blanket proclamations is wrong. Today @ performs
> better than @: most of the time, especially for primitives. But for defined
> verbs, depending on the design, there may be no performance gain and a lot
> of wasted memory.
>
> If you have an app with a performance problem then run a tool to find out
> where you're spending the time. J has such a tool, as do almost all
> programming languages. Now you know where to spend your time.
>
> And you may find out that the problem is not in J, but in your design.
>
> There is no argument that @ has much room for improvement. Which seems to
> be what your beef is about. So you're wasting your time and everybody
> else's time proving the obvious.
>
> On Oct 2, 2017 10:06 AM, "Erling Hellenäs"  wrote:
>
>> You are welcome to show us better measurements if you think these
>> measurements are very important and worthy of a lengthy discussion in the
>> forum. /Erling
>>
>> On 2017-10-02 17:53, Raul Miller wrote:
>>
>>> The null case here should be ]
>>>
>>> (-@- -: ]) v
>>> 1
>>>
>>> Thanks,
>>>
>>>
>> --
>> 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

Re: [Jprogramming] advancing in labs in jconsole

[chris burke]

In j806 console, to advance enter:

   lab 0



On Mon, Oct 2, 2017 at 10:18 AM, 'Pascal Jasmin' via Programming <
programm...@jsoftware.com> wrote:

> How would I advance to next section in a lab with jconsole 8.06 (windows
> 10)?
>
> |  | Virus-free. www.avast.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

Re: [Jprogramming] Composition conjunctions

[Don Guinn]

Detailed measurements are useful and meaningful if your application is
performing badly. But general statements about poor performance in parts of
an application that isn't used much is a waste of time.

@ vs. @: is a concern but blanket proclamations is wrong. Today @ performs
better than @: most of the time, especially for primitives. But for defined
verbs, depending on the design, there may be no performance gain and a lot
of wasted memory.

If you have an app with a performance problem then run a tool to find out
where you're spending the time. J has such a tool, as do almost all
programming languages. Now you know where to spend your time.

And you may find out that the problem is not in J, but in your design.

There is no argument that @ has much room for improvement. Which seems to
be what your beef is about. So you're wasting your time and everybody
else's time proving the obvious.

On Oct 2, 2017 10:06 AM, "Erling Hellenäs"  wrote:

> You are welcome to show us better measurements if you think these
> measurements are very important and worthy of a lengthy discussion in the
> forum. /Erling
>
> On 2017-10-02 17:53, Raul Miller wrote:
>
>> The null case here should be ]
>>
>> (-@- -: ]) v
>> 1
>>
>> Thanks,
>>
>>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

[Jprogramming] advancing in labs in jconsole

['Pascal Jasmin' via Programming]

How would I advance to next section in a lab with jconsole 8.06 (windows 10)?

|  | Virus-free. www.avast.com  |

--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] (no subject)

[Marc Simpson]

More naive than Raul's approach, first pass using the 'stats' lib:

   a=.2 5 7
   require'stats'
   combv=: ] {~ (comb #@])
   3 combv a
2 5 7
   2 combv a
2 5
2 7
5 7
   1 combv a
2
5
7
   combos=: (1 + i.@#) <@combv"0 1 ]
   combos a
┌─┬───┬─┐
│2│2 5│2 5 7│
│5│2 7│ │
│7│5 7│ │
└─┴───┴─┘
   f=: 1 : ';u/"1 each combos y'
   +f a
2 5 7 7 9 12 14
   *f a
2 5 7 10 14 35 70

/M

On Mon, Oct 2, 2017 at 10:06 AM, Raul Miller  wrote:
> For a first effort, I would go with
>
>   F=:1 :'(u@#~ #:@i.@(2^#))@~.'
>   f=: /F
>
> Hopefully that makes the issues obvious - the specification here calls
> for a result which grows exponentially with the size of the argument
> set.
>
> Also:
>
> The ~. might be extra work, but for typical cases the effort of
> ensuring that the argument is a set is trivial compared to the effort
> of constructing the result.
>
> You did not include the empty combination in your example results, but
> given your specification my initial inclination is to treat that as an
> oversight.
>
> I defined F instead of going straight for f because for testing
> purposes I want to be able to do (
> Thanks,
>
> --
> Raul
>
>
> On Mon, Oct 2, 2017 at 12:49 PM, Skip Cave  wrote:
>> Given a set of integers, what is the most concise and or efficient way to
>> list the numbers along with the sum of all combinations of the numbers? the
>> products of all combinations?
>>
>> for example:
>>
>> a =. 2 5 7
>> + f aNB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
>> 2 5 7 7 9 12 14
>>
>> * f aNB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
>> 2 5 7 10 14 35 70
>>
>> The function 'f' should work for any verb and any size right argument noun
>> vector.
>>
>> Skip
>>
>> Skip Cave
>> Cave Consulting LLC
>> --
>> 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

Re: [Jprogramming] (no subject)

[Raul Miller]

For a first effort, I would go with

  F=:1 :'(u@#~ #:@i.@(2^#))@~.'
  f=: /F

Hopefully that makes the issues obvious - the specification here calls
for a result which grows exponentially with the size of the argument
set.

Also:

The ~. might be extra work, but for typical cases the effort of
ensuring that the argument is a set is trivial compared to the effort
of constructing the result.

You did not include the empty combination in your example results, but
given your specification my initial inclination is to treat that as an
oversight.

I defined F instead of going straight for f because for testing
purposes I want to be able to do ( wrote:
> Given a set of integers, what is the most concise and or efficient way to
> list the numbers along with the sum of all combinations of the numbers? the
> products of all combinations?
>
> for example:
>
> a =. 2 5 7
> + f aNB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
> 2 5 7 7 9 12 14
>
> * f aNB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
> 2 5 7 10 14 35 70
>
> The function 'f' should work for any verb and any size right argument noun
> vector.
>
> Skip
>
> Skip Cave
> Cave Consulting LLC
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

[Jprogramming] (no subject)

[Skip Cave]

Given a set of integers, what is the most concise and or efficient way to
list the numbers along with the sum of all combinations of the numbers? the
products of all combinations?

for example:

a =. 2 5 7
+ f aNB. 2, 5, 7, (2+5), (2+7), (5+7), (2+5+7)
2 5 7 7 9 12 14

* f aNB. 2, 5, 7, (2*5), (2*7), (5*7), (2*5*7)
2 5 7 10 14 35 70

The function 'f' should work for any verb and any size right argument noun
vector.

Skip

Skip Cave
Cave Consulting LLC
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Raul Miller]

I actually don't think these particular measurements are all that
relevant. And I guess I should apologize for not making that clear.

That said, I think timing the test suite which comes with j source
(except for the *t.* files -- too many race conditions there) is kind
of interesting.

Thanks,

-- 
Raul

On Mon, Oct 2, 2017 at 12:06 PM, Erling Hellenäs
 wrote:
> You are welcome to show us better measurements if you think these
> measurements are very important and worthy of a lengthy discussion in the
> forum. /Erling
>
> On 2017-10-02 17:53, Raul Miller wrote:
>>
>> The null case here should be ]
>>
>> (-@- -: ]) v
>> 1
>>
>> Thanks,
>>
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

You are welcome to show us better measurements if you think these 
measurements are very important and worthy of a lengthy discussion in 
the forum. /Erling


On 2017-10-02 17:53, Raul Miller wrote:

The null case here should be ]

(-@- -: ]) v
1

Thanks,



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Hi all !

Here are the same measurements with Bill Lams proposed way to measure 
this. As we can all see it is lousy because the overhead in the 
measurement loop is much higher than that in @.


   n=.50
   v=.?~n
   $v
50
   10{.v
414253 208033 396027 190208 112248 196018 51503 465441 277514 114236
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.258554 1.32199e8
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.260208 1.322e8
   ts'-At-At-At-At-Atop- v'
2.98929 1.32204e8
   ts'-@:-@:-@:-@:-@- v'
0.154225 1.32197e8
   ts'-@- v'
0.0907723 1.32195e8
   v=.1
   n*{.n ts'-@:-@:-@:-@:- v'
0.455432

Cheers,

Erling


On 2017-10-02 17:24, Raul Miller wrote:

Believe me, I understood that.

But this particular performance measure is rather like measuring the
performance of tires sliding sideways (as opposed to rolling) in a
parking lot while carrying several hundred pounds of meatloaf.

In other words, it's not something I would feel comfortable optimizing
for, though I can sort of almost see some sort of vague connection to
real applications if I do not think about it too much.

Thanks,



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Raul Miller]

The null case here should be ]

   (-@- -: ]) v
1

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 11:48 AM, Erling Hellenäs
 wrote:
> Hi all !
>
> Here are the same measurements with a null case we can deduct, if we want to
> know the exact speed on my particular machine.
>
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
> 0.261088 1.32199e8
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
> 0.264798 1.322e8
>ts'-At-At-At-At-Atop- v'
> 2.97941 1.32204e8
>ts'-@:-@:-@:-@:-@- v'
> 0.156113 1.32197e8
>ts'-@- v'
> 0.091478 1.32195e8
>
> Cheers,
> Erling
>
> On 2017-10-02 17:24, Raul Miller wrote:
>>
>> Believe me, I understood that.
>>
>> But this particular performance measure is rather like measuring the
>> performance of tires sliding sideways (as opposed to rolling) in a
>> parking lot while carrying several hundred pounds of meatloaf.
>>
>> In other words, it's not something I would feel comfortable optimizing
>> for, though I can sort of almost see some sort of vague connection to
>> real applications if I do not think about it too much.
>>
>> Thanks,
>>
>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Hi all !

Here are the same measurements with a null case we can deduct, if we 
want to know the exact speed on my particular machine.


   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.261088 1.32199e8
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.264798 1.322e8
   ts'-At-At-At-At-Atop- v'
2.97941 1.32204e8
   ts'-@:-@:-@:-@:-@- v'
0.156113 1.32197e8
   ts'-@- v'
0.091478 1.32195e8

Cheers,
Erling

On 2017-10-02 17:24, Raul Miller wrote:

Believe me, I understood that.

But this particular performance measure is rather like measuring the
performance of tires sliding sideways (as opposed to rolling) in a
parking lot while carrying several hundred pounds of meatloaf.

In other words, it's not something I would feel comfortable optimizing
for, though I can sort of almost see some sort of vague connection to
real applications if I do not think about it too much.

Thanks,



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Well, I do some observations and if anyone finds them interesting he can 
respond. Anyone can do their own measurement. It does not obviously 
improve the discussion to look for all kinds of perceived faults in 
peoples posts and start discussion these, instead of the subject of the 
thread? /Erling


On 2017-10-02 17:24, Raul Miller wrote:

Believe me, I understood that.

But this particular performance measure is rather like measuring the
performance of tires sliding sideways (as opposed to rolling) in a
parking lot while carrying several hundred pounds of meatloaf.

In other words, it's not something I would feel comfortable optimizing
for, though I can sort of almost see some sort of vague connection to
real applications if I do not think about it too much.

Thanks,



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Yes, I was not sure if it might only concern the Unix binaries, but 
since it worked on my machine it seemed to concern the Windows binaries 
too. Thanks for clearing it out.  /Erling


On 2017-10-02 17:26, Jose Mario Quintana wrote:

" I guess the Jx j.dll is compiled without it."

That is correct, both, the provided dll and the so binaries do not support
avx as stated in the release post:

"

Jx 1.1 Release

A Jx v1.1 Extensions Guide, a J/Jx Cheatsheet, a Jx Assertions script
together with links to a Windows 64 bit dll, a Unix 64 bit so binaries
(without avx support) and the patch corresponding to the J806 source
(beta-6) can be found at the link [0].

"


On Mon, Oct 2, 2017 at 10:11 AM, Erling Hellenäs 
wrote:


Hi all !

These measurements are with Jx j.dll on the latest Beta release.

n=.50
v=.?~n
$v
50
10{.v
50689 85238 136693 442735 25192 304048 139323 96663 216893 142946
ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.1156 4.19917e6
ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.108927 4.20006e6
ts'-At-At-At-At-Atop- v'
3.30635 4.20416e6
ts'-@:-@:-@:-@:-@- v'
0.115168 4.19712e6

For some reason my tacit code is on par with the built in code here. The
performance of my tacit code is doubled!!! in this release compared to
latest stable. There is no avx on this machine. I guess the Jx j.dll is
compiled without it. The explicit code is even slower here.

The measurements from the latest stable, taken from below:

n=.50
v=.?~n
$v
50
10{.v
415593 299586 376558 161399 308885 477430 286004 354448 123594 75795
ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.240852 1.32199e8
ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.2369 1.322e8
ts'-At-At-At-At-Atop- v'
2.96161 1.32204e8
ts'-@:-@:-@:-@:-@- v'
0.158078 1.32197e8

Cheers,

Erling


On 2017-10-02 15:42, Erling Hellenäs wrote:


Hi all !

I tried to create explicit and tacit definitions of our composition
conjunctions. I think the result could possibly be used to clarify the
descriptions in NuVoc.

Opinions are welcome, there are probably still some bugs or
misunderstandings of function, there could be interesting aspects to
discuss.

The printout follows, then my project with definitions, tests and some
minor explanations.

In the end of the printout there are some performance measurements. The
explicit versions are doing bad in those. The built in versions are
slightly faster, which could be expected.

Cheers,

Erling

Printout===

ts=: 6!:2 , 7!:2@]   NB. Time and space

At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

 1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfinite> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
NB. < 2 2 $ 4 3 3 2

NB. &.:
UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

-UnderRankInfiniteTacitMonadic> 1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
NB. < 2 2 $ 4 3 3 2

NB. &.:
-UnderRankInfiniteTacitMonadic>
[: >^:_1 [: - [: > ]
-UnderRankInfiniteTacitDyadic>
[: >^:_1 ([: > [) - [: > ]

UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u UnderRankInfinite v"v y
)

-UnderRankV> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
NB. _1 _2; _3 _4
(5;6) -UnderRankV> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
NB. 4 3;3 2

NB. &.
UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'

-UnderRankVTacitMonadic> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
NB. _1 _2; _3 _4
(5;6) -UnderRankVTacitDyadic> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
NB. 4 3;3 2

NB. &.
-UnderRankVTacitMonadic>
([: >^:_1 [: - [: > ])"0 0 0
-UnderRankVTacitDyadic>
([: >^:_1 ([: > [) - [: > ])"0 0 0

n=.50
v=.?~n
$v
50
10{.v
415593 299586 376558 161399 308885 477430 286004 354448 123594 75795

ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.240852 1.32199e8
ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.2369 1.322e8
ts'-At-At-At-At-Atop- v'
2.96161 1.32204e8
ts'-@:-@:-@:-@:-@- v'
0.158078 1.32197e8

== Project ===

ts=: 6!:2 , 7!:2@]   NB. Time and space

At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfinite> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

-UnderRankInfiniteTacitMonadic> 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
-UnderRankInfiniteTacitMonadic>
-UnderRankInfiniteTacitDyadic>

UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u 

Re: [Jprogramming] Composition conjunctions

[Jose Mario Quintana]

" I guess the Jx j.dll is compiled without it."

That is correct, both, the provided dll and the so binaries do not support
avx as stated in the release post:

"

Jx 1.1 Release

A Jx v1.1 Extensions Guide, a J/Jx Cheatsheet, a Jx Assertions script
together with links to a Windows 64 bit dll, a Unix 64 bit so binaries
(without avx support) and the patch corresponding to the J806 source
(beta-6) can be found at the link [0].

"


On Mon, Oct 2, 2017 at 10:11 AM, Erling Hellenäs 
wrote:

> Hi all !
>
> These measurements are with Jx j.dll on the latest Beta release.
>
>n=.50
>v=.?~n
>$v
> 50
>10{.v
> 50689 85238 136693 442735 25192 304048 139323 96663 216893 142946
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
> 0.1156 4.19917e6
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
> 0.108927 4.20006e6
>ts'-At-At-At-At-Atop- v'
> 3.30635 4.20416e6
>ts'-@:-@:-@:-@:-@- v'
> 0.115168 4.19712e6
>
> For some reason my tacit code is on par with the built in code here. The
> performance of my tacit code is doubled!!! in this release compared to
> latest stable. There is no avx on this machine. I guess the Jx j.dll is
> compiled without it. The explicit code is even slower here.
>
> The measurements from the latest stable, taken from below:
>
>n=.50
>v=.?~n
>$v
> 50
>10{.v
> 415593 299586 376558 161399 308885 477430 286004 354448 123594 75795
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
> 0.240852 1.32199e8
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
> 0.2369 1.322e8
>ts'-At-At-At-At-Atop- v'
> 2.96161 1.32204e8
>ts'-@:-@:-@:-@:-@- v'
> 0.158078 1.32197e8
>
> Cheers,
>
> Erling
>
>
> On 2017-10-02 15:42, Erling Hellenäs wrote:
>
>> Hi all !
>>
>> I tried to create explicit and tacit definitions of our composition
>> conjunctions. I think the result could possibly be used to clarify the
>> descriptions in NuVoc.
>>
>> Opinions are welcome, there are probably still some bugs or
>> misunderstandings of function, there could be interesting aspects to
>> discuss.
>>
>> The printout follows, then my project with definitions, tests and some
>> minor explanations.
>>
>> In the end of the printout there are some performance measurements. The
>> explicit versions are doing bad in those. The built in versions are
>> slightly faster, which could be expected.
>>
>> Cheers,
>>
>> Erling
>>
>> Printout===
>>
>>ts=: 6!:2 , 7!:2@]   NB. Time and space
>>
>>At=: 2 : 0
>> NB. @:
>> u v"_ y
>> :
>> u x v"_ y
>> )
>>
>>> ┌─┐
>> │_1 _2│
>> └─┘
>>NB. < _1_2
>>5 > ┌───┐
>> │4 3│
>> └───┘
>>NB. < 4 3
>>
>>NB. @:
>>AtTacit=: 2 : '[: u v'
>>
>>
>>> ┌─┐
>> │_1 _2│
>> └─┘
>>NB. < _1_2
>>5 > ┌───┐
>> │4 3│
>> └───┘
>>NB. < 4 3
>>
>>NB. @:
>>> [: < -
>>
>>Atop=: 2 : 0
>> NB. @
>> u At v"v y
>> :
>> x u At v"v y
>> )
>>
>>> ┌──┬──┐
>> │_1│_2│
>> └──┴──┘
>>NB. _1;_2
>>5 > ┌─┬─┐
>> │4│3│
>> └─┴─┘
>>NB. 4;3
>>
>>NB. @
>>AtopTacit=: 2 : '([: u v)"v'
>>
>>> ┌──┬──┐
>> │_1│_2│
>> └──┴──┘
>>NB. _1;_2
>>5 > ┌─┬─┐
>> │4│3│
>> └─┴─┘
>>NB. 4;3
>>
>>NB. @
>>> ([: < -)"0 0 0
>>> ([: < [)"_ _ _
>>> ([: < i.)"1 _ _
>>
>>Appose=: 2 : 0
>> NB. &:
>> u v"_ y
>> :
>> (v"_ x) u v"_ y
>> )
>>
>>> ┌─┐
>> │_1 _2│
>> └─┘
>>NB. < _1 _2
>>5 ([:<<)Appose- 4 5
>> ┌───┐
>> │1 0│
>> └───┘
>>NB. < 1 0
>>
>>NB. &:
>>ApposeTacitMonadic=: 2 : '[: u [: v ]'
>>ApposeTacitDyadic=: 2 : '([: v [) u [: v ]'
>>
>>> ┌─┐
>> │_1 _2│
>> └─┘
>>NB. < _1 _2
>>5 ([:<<)ApposeTacitDyadic- 4 5
>> ┌───┐
>> │1 0│
>> └───┘
>>NB. < 1 0
>>
>>NB. &:
>>> [: < [: - ]
>>([:<<)ApposeTacitDyadic-
>> ([: - [) ([: < <) [: - ]
>>
>>Compose=: 2 : 0
>> NB. &
>> u Appose v"v y
>> :
>> x u Appose v"v y
>> )
>>
>>> ┌──┬──┐
>> │_1│_2│
>> └──┴──┘
>>NB. _1;_2
>>5 ([:<<)Compose- 4 5
>> ┌─┬─┐
>> │1│0│
>> └─┴─┘
>>NB. 1;0
>>
>>NB. &
>>ComposeTacitMonadic=: 2 : '([: u [: v ])"v'
>>ComposeTacitDyadic=: 2 : '(([: v [) u [: v ])"v'
>>
>>> ┌──┬──┐
>> │_1│_2│
>> └──┴──┘
>>NB. _1;_2
>>5 ([:<<)ComposeTacitDyadic- 4 5
>> ┌─┬─┐
>> │1│0│
>> └─┴─┘
>>NB. 1;0
>>
>>NB. &
>>> ([: < [: - ])"0 0 0
>>([:<<)ComposeTacitDyadic-
>> (([: - [) ([: < <) [: - ])"0 0 0
>>
>>UnderRankInfinite=: 2 : 0
>> NB. &.:
>> v"_^:_1 u v"_ y
>> :
>> v"_^:_1 (v"_ x) u v"_ y
>> )
>>
>>-UnderRankInfinite> 1 2;3 4
>> ┌─┐
>> │_1 _2│
>> │_3 _4│
>> └─┘
>>NB. <2 2 $ _1 _2 -3 -4
>>(5;6) -UnderRankInfinite> 1 2;3 4
>> ┌───┐
>> │4 3│
>> │3 2│
>> └───┘
>>NB. < 2 2 $ 4 3 3 2
>>
>>NB. &.:
>>UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
>>UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'
>>
>>-UnderRankInfiniteTacitMonadic> 1 2;3 4
>> ┌─┐
>> │_1 _2│
>> │_3 _4│
>> └─┘
>>NB. <2 2 $ _1 _2 -3 

Re: [Jprogramming] Composition conjunctions

[Raul Miller]

Believe me, I understood that.

But this particular performance measure is rather like measuring the
performance of tires sliding sideways (as opposed to rolling) in a
parking lot while carrying several hundred pounds of meatloaf.

In other words, it's not something I would feel comfortable optimizing
for, though I can sort of almost see some sort of vague connection to
real applications if I do not think about it too much.

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 11:12 AM, Erling Hellenäs
 wrote:
> Even  if Raul wouldn't consider them meaningful for some reason I don't
> understand it seems he should still understand that they are performance
> measurements? That I am creating a loop over a piece of code to measure the
> performance of this piece of code?  /Erling
>
>
> On 2017-10-02 17:01, Don Guinn wrote:
>>
>> How many times do you have to be told that these kinds of measurements and
>> meaningless?
>>
>> On Oct 2, 2017 8:57 AM, "Erling Hellenäs" 
>> wrote:
>>
>>> How many times do I have to tell that I do this to measure performance?
>>> /Erling
>>>
>>> On 2017-10-02 16:39, Raul Miller wrote:
>>>
 Generally speaking, you want to push the large arrays into J's
 primitives as much as possible. Going the other direction, like you're
 doing here (using relatively expensive functions at rank 0 on
 relatively large arrays) is mostly going to bog down.

 Thanks,


>>> --
>>> 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
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Even  if Raul wouldn't consider them meaningful for some reason I don't 
understand it seems he should still understand that they are performance 
measurements? That I am creating a loop over a piece of code to measure 
the performance of this piece of code?  /Erling


On 2017-10-02 17:01, Don Guinn wrote:

How many times do you have to be told that these kinds of measurements and
meaningless?

On Oct 2, 2017 8:57 AM, "Erling Hellenäs"  wrote:


How many times do I have to tell that I do this to measure performance?
/Erling

On 2017-10-02 16:39, Raul Miller wrote:


Generally speaking, you want to push the large arrays into J's
primitives as much as possible. Going the other direction, like you're
doing here (using relatively expensive functions at rank 0 on
relatively large arrays) is mostly going to bog down.

Thanks,



--
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

Re: [Jprogramming] Composition conjunctions

[Don Guinn]

How many times do you have to be told that these kinds of measurements and
meaningless?

On Oct 2, 2017 8:57 AM, "Erling Hellenäs"  wrote:

> How many times do I have to tell that I do this to measure performance?
> /Erling
>
> On 2017-10-02 16:39, Raul Miller wrote:
>
>> Generally speaking, you want to push the large arrays into J's
>> primitives as much as possible. Going the other direction, like you're
>> doing here (using relatively expensive functions at rank 0 on
>> relatively large arrays) is mostly going to bog down.
>>
>> Thanks,
>>
>>
> --
> For information about J forums see http://www.jsoftware.com/forums.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

How many times do I have to tell that I do this to measure performance? 
/Erling


On 2017-10-02 16:39, Raul Miller wrote:

Generally speaking, you want to push the large arrays into J's
primitives as much as possible. Going the other direction, like you're
doing here (using relatively expensive functions at rank 0 on
relatively large arrays) is mostly going to bog down.

Thanks,



--
For information about J forums see http://www.jsoftware.com/forums.htm

Re: [Jprogramming] Composition conjunctions

[Raul Miller]

Generally speaking, you want to push the large arrays into J's
primitives as much as possible. Going the other direction, like you're
doing here (using relatively expensive functions at rank 0 on
relatively large arrays) is mostly going to bog down.

Thanks,

-- 
Raul


On Mon, Oct 2, 2017 at 9:42 AM, Erling Hellenäs
 wrote:
> Hi all !
>
> I tried to create explicit and tacit definitions of our composition
> conjunctions. I think the result could possibly be used to clarify the
> descriptions in NuVoc.
>
> Opinions are welcome, there are probably still some bugs or
> misunderstandings of function, there could be interesting aspects to
> discuss.
>
> The printout follows, then my project with definitions, tests and some minor
> explanations.
>
> In the end of the printout there are some performance measurements. The
> explicit versions are doing bad in those. The built in versions are slightly
> faster, which could be expected.
>
> Cheers,
>
> Erling
>
> Printout===
>
>ts=: 6!:2 , 7!:2@]   NB. Time and space
>
>At=: 2 : 0
> NB. @:
> u v"_ y
> :
> u x v"_ y
> )
>
> ┌─┐
> │_1 _2│
> └─┘
>NB. < _1_2
>5  ┌───┐
> │4 3│
> └───┘
>NB. < 4 3
>
>NB. @:
>AtTacit=: 2 : '[: u v'
>
>
> ┌─┐
> │_1 _2│
> └─┘
>NB. < _1_2
>5  ┌───┐
> │4 3│
> └───┘
>NB. < 4 3
>
>NB. @:
> [: < -
>
>Atop=: 2 : 0
> NB. @
> u At v"v y
> :
> x u At v"v y
> )
>
> ┌──┬──┐
> │_1│_2│
> └──┴──┘
>NB. _1;_2
>5  ┌─┬─┐
> │4│3│
> └─┴─┘
>NB. 4;3
>
>NB. @
>AtopTacit=: 2 : '([: u v)"v'
>
> ┌──┬──┐
> │_1│_2│
> └──┴──┘
>NB. _1;_2
>5  ┌─┬─┐
> │4│3│
> └─┴─┘
>NB. 4;3
>
>NB. @
> ([: < -)"0 0 0
> ([: < [)"_ _ _
> ([: < i.)"1 _ _
>
>Appose=: 2 : 0
> NB. &:
> u v"_ y
> :
> (v"_ x) u v"_ y
> )
>
> ┌─┐
> │_1 _2│
> └─┘
>NB. < _1 _2
>5 ([:<<)Appose- 4 5
> ┌───┐
> │1 0│
> └───┘
>NB. < 1 0
>
>NB. &:
>ApposeTacitMonadic=: 2 : '[: u [: v ]'
>ApposeTacitDyadic=: 2 : '([: v [) u [: v ]'
>
> ┌─┐
> │_1 _2│
> └─┘
>NB. < _1 _2
>5 ([:<<)ApposeTacitDyadic- 4 5
> ┌───┐
> │1 0│
> └───┘
>NB. < 1 0
>
>NB. &:
> [: < [: - ]
>([:<<)ApposeTacitDyadic-
> ([: - [) ([: < <) [: - ]
>
>Compose=: 2 : 0
> NB. &
> u Appose v"v y
> :
> x u Appose v"v y
> )
>
> ┌──┬──┐
> │_1│_2│
> └──┴──┘
>NB. _1;_2
>5 ([:<<)Compose- 4 5
> ┌─┬─┐
> │1│0│
> └─┴─┘
>NB. 1;0
>
>NB. &
>ComposeTacitMonadic=: 2 : '([: u [: v ])"v'
>ComposeTacitDyadic=: 2 : '(([: v [) u [: v ])"v'
>
> ┌──┬──┐
> │_1│_2│
> └──┴──┘
>NB. _1;_2
>5 ([:<<)ComposeTacitDyadic- 4 5
> ┌─┬─┐
> │1│0│
> └─┴─┘
>NB. 1;0
>
>NB. &
> ([: < [: - ])"0 0 0
>([:<<)ComposeTacitDyadic-
> (([: - [) ([: < <) [: - ])"0 0 0
>
>UnderRankInfinite=: 2 : 0
> NB. &.:
> v"_^:_1 u v"_ y
> :
> v"_^:_1 (v"_ x) u v"_ y
> )
>
>-UnderRankInfinite> 1 2;3 4
> ┌─┐
> │_1 _2│
> │_3 _4│
> └─┘
>NB. <2 2 $ _1 _2 -3 -4
>(5;6) -UnderRankInfinite> 1 2;3 4
> ┌───┐
> │4 3│
> │3 2│
> └───┘
>NB. < 2 2 $ 4 3 3 2
>
>NB. &.:
>UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
>UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'
>
>-UnderRankInfiniteTacitMonadic> 1 2;3 4
> ┌─┐
> │_1 _2│
> │_3 _4│
> └─┘
>NB. <2 2 $ _1 _2 -3 -4
>(5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
> ┌───┐
> │4 3│
> │3 2│
> └───┘
>NB. < 2 2 $ 4 3 3 2
>
>NB. &.:
>-UnderRankInfiniteTacitMonadic>
> [: >^:_1 [: - [: > ]
>-UnderRankInfiniteTacitDyadic>
> [: >^:_1 ([: > [) - [: > ]
>
>UnderRankV=: 2 : 0
> NB. &.
> u UnderRankInfinite v"v y
> :
> x u UnderRankInfinite v"v y
> )
>
>-UnderRankV> 1 2; 3 4
> ┌─┬─┐
> │_1 _2│_3 _4│
> └─┴─┘
>NB. _1 _2; _3 _4
>(5;6) -UnderRankV> 1 2;3 4
> ┌───┬───┐
> │4 3│3 2│
> └───┴───┘
>NB. 4 3;3 2
>
>NB. &.
>UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
>UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'
>
>-UnderRankVTacitMonadic> 1 2; 3 4
> ┌─┬─┐
> │_1 _2│_3 _4│
> └─┴─┘
>NB. _1 _2; _3 _4
>(5;6) -UnderRankVTacitDyadic> 1 2;3 4
> ┌───┬───┐
> │4 3│3 2│
> └───┴───┘
>NB. 4 3;3 2
>
>NB. &.
>-UnderRankVTacitMonadic>
> ([: >^:_1 [: - [: > ])"0 0 0
>-UnderRankVTacitDyadic>
> ([: >^:_1 ([: > [) - [: > ])"0 0 0
>
>n=.50
>v=.?~n
>$v
> 50
>10{.v
> 415593 299586 376558 161399 308885 477430 286004 354448 123594 75795
>
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
> 0.240852 1.32199e8
>ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
> 0.2369 1.322e8
>ts'-At-At-At-At-Atop- v'
> 2.96161 1.32204e8
>ts'-@:-@:-@:-@:-@- v'
> 0.158078 1.32197e8
>
> == Project ===
>
> ts=: 6!:2 , 7!:2@]   NB. Time and space
>
> At=: 2 : 0
> NB. @:
> u v"_ y
> :
> u x v"_ y
> )
>
>  NB. < _1_2
> 5  NB. < 4 3
>
> NB. 

Re: [Jprogramming] Composition conjunctions

[Erling Hellenäs]

Hi all !

These measurements are with Jx j.dll on the latest Beta release.

   n=.50
   v=.?~n
   $v
50
   10{.v
50689 85238 136693 442735 25192 304048 139323 96663 216893 142946
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.1156 4.19917e6
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.108927 4.20006e6
   ts'-At-At-At-At-Atop- v'
3.30635 4.20416e6
   ts'-@:-@:-@:-@:-@- v'
0.115168 4.19712e6

For some reason my tacit code is on par with the built in code here. The 
performance of my tacit code is doubled!!! in this release compared to 
latest stable. There is no avx on this machine. I guess the Jx j.dll is 
compiled without it. The explicit code is even slower here.


The measurements from the latest stable, taken from below:

   n=.50
   v=.?~n
   $v
50
   10{.v
415593 299586 376558 161399 308885 477430 286004 354448 123594 75795
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.240852 1.32199e8
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.2369 1.322e8
   ts'-At-At-At-At-Atop- v'
2.96161 1.32204e8
   ts'-@:-@:-@:-@:-@- v'
0.158078 1.32197e8

Cheers,

Erling

On 2017-10-02 15:42, Erling Hellenäs wrote:

Hi all !

I tried to create explicit and tacit definitions of our composition 
conjunctions. I think the result could possibly be used to clarify the 
descriptions in NuVoc.


Opinions are welcome, there are probably still some bugs or 
misunderstandings of function, there could be interesting aspects to 
discuss.


The printout follows, then my project with definitions, tests and some 
minor explanations.


In the end of the printout there are some performance measurements. 
The explicit versions are doing bad in those. The built in versions 
are slightly faster, which could be expected.


Cheers,

Erling

Printout===

   ts=: 6!:2 , 7!:2@]   NB. Time and space

   At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

    1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
   NB. <2 2 $ _1 _2 -3 -4
   (5;6) -UnderRankInfinite> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
   NB. < 2 2 $ 4 3 3 2

   NB. &.:
   UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
   UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

   -UnderRankInfiniteTacitMonadic> 1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
   NB. <2 2 $ _1 _2 -3 -4
   (5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
   NB. < 2 2 $ 4 3 3 2

   NB. &.:
   -UnderRankInfiniteTacitMonadic>
[: >^:_1 [: - [: > ]
   -UnderRankInfiniteTacitDyadic>
[: >^:_1 ([: > [) - [: > ]

   UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u UnderRankInfinite v"v y
)

   -UnderRankV> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
   NB. _1 _2; _3 _4
   (5;6) -UnderRankV> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
   NB. 4 3;3 2

   NB. &.
   UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
   UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'

   -UnderRankVTacitMonadic> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
   NB. _1 _2; _3 _4
   (5;6) -UnderRankVTacitDyadic> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
   NB. 4 3;3 2

   NB. &.
   -UnderRankVTacitMonadic>
([: >^:_1 [: - [: > ])"0 0 0
   -UnderRankVTacitDyadic>
([: >^:_1 ([: > [) - [: > ])"0 0 0

   n=.50
   v=.?~n
   $v
50
   10{.v
415593 299586 376558 161399 308885 477430 286004 354448 123594 75795

   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.240852 1.32199e8
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.2369 1.322e8
   ts'-At-At-At-At-Atop- v'
2.96161 1.32204e8
   ts'-@:-@:-@:-@:-@- v'
0.158078 1.32197e8

== Project ===

ts=: 6!:2 , 7!:2@]   NB. Time and space

At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfinite> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

-UnderRankInfiniteTacitMonadic> 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
-UnderRankInfiniteTacitMonadic>
-UnderRankInfiniteTacitDyadic>

UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u UnderRankInfinite v"v y
)

-UnderRankV> 1 2; 3 4
NB. _1 _2; _3 _4
(5;6) -UnderRankV> 1 2;3 4
NB. 4 3;3 2

NB. &.
UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'

-UnderRankVTacitMonadic> 1 2; 3 4
NB. _1 _2; _3 _4
(5;6) -UnderRankVTacitDyadic> 1 2;3 4
NB. 4 3;3 2

NB. &.
-UnderRankVTacitMonadic>
-UnderRankVTacitDyadic>

n=.50
v=.?~n
$v
10{.v

ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
ts'-At-At-At-At-Atop- v'
ts'-@:-@:-@:-@:-@- v'

--
For information about J forums see http://www.jsoftware.com/forums.htm



--
For information about J forums 

[Jprogramming] Composition conjunctions

[Erling Hellenäs]

Hi all !

I tried to create explicit and tacit definitions of our composition 
conjunctions. I think the result could possibly be used to clarify the 
descriptions in NuVoc.


Opinions are welcome, there are probably still some bugs or 
misunderstandings of function, there could be interesting aspects to 
discuss.


The printout follows, then my project with definitions, tests and some 
minor explanations.


In the end of the printout there are some performance measurements. The 
explicit versions are doing bad in those. The built in versions are 
slightly faster, which could be expected.


Cheers,

Erling

Printout===

   ts=: 6!:2 , 7!:2@]   NB. Time and space

   At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

    1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
   NB. <2 2 $ _1 _2 -3 -4
   (5;6) -UnderRankInfinite> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
   NB. < 2 2 $ 4 3 3 2

   NB. &.:
   UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
   UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

   -UnderRankInfiniteTacitMonadic> 1 2;3 4
┌─┐
│_1 _2│
│_3 _4│
└─┘
   NB. <2 2 $ _1 _2 -3 -4
   (5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
┌───┐
│4 3│
│3 2│
└───┘
   NB. < 2 2 $ 4 3 3 2

   NB. &.:
   -UnderRankInfiniteTacitMonadic>
[: >^:_1 [: - [: > ]
   -UnderRankInfiniteTacitDyadic>
[: >^:_1 ([: > [) - [: > ]

   UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u UnderRankInfinite v"v y
)

   -UnderRankV> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
   NB. _1 _2; _3 _4
   (5;6) -UnderRankV> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
   NB. 4 3;3 2

   NB. &.
   UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
   UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'

   -UnderRankVTacitMonadic> 1 2; 3 4
┌─┬─┐
│_1 _2│_3 _4│
└─┴─┘
   NB. _1 _2; _3 _4
   (5;6) -UnderRankVTacitDyadic> 1 2;3 4
┌───┬───┐
│4 3│3 2│
└───┴───┘
   NB. 4 3;3 2

   NB. &.
   -UnderRankVTacitMonadic>
([: >^:_1 [: - [: > ])"0 0 0
   -UnderRankVTacitDyadic>
([: >^:_1 ([: > [) - [: > ])"0 0 0

   n=.50
   v=.?~n
   $v
50
   10{.v
415593 299586 376558 161399 308885 477430 286004 354448 123594 75795

   ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
0.240852 1.32199e8
   ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
0.2369 1.322e8
   ts'-At-At-At-At-Atop- v'
2.96161 1.32204e8
   ts'-@:-@:-@:-@:-@- v'
0.158078 1.32197e8

== Project ===

ts=: 6!:2 , 7!:2@]   NB. Time and space

At=: 2 : 0
NB. @:
u v"_ y
:
u x v"_ y
)

 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfinite> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
UnderRankInfiniteTacitMonadic=: 2 : '[: v^:_1 [: u [: v ]'
UnderRankInfiniteTacitDyadic=: 2 : '[: v^:_1 ([: v [) u [: v ]'

-UnderRankInfiniteTacitMonadic> 1 2;3 4
NB. <2 2 $ _1 _2 -3 -4
(5;6) -UnderRankInfiniteTacitDyadic> 1 2;3 4
NB. < 2 2 $ 4 3 3 2

NB. &.:
-UnderRankInfiniteTacitMonadic>
-UnderRankInfiniteTacitDyadic>

UnderRankV=: 2 : 0
NB. &.
u UnderRankInfinite v"v y
:
x u UnderRankInfinite v"v y
)

-UnderRankV> 1 2; 3 4
NB. _1 _2; _3 _4
(5;6) -UnderRankV> 1 2;3 4
NB. 4 3;3 2

NB. &.
UnderRankVTacitMonadic=: 2 : '([: v^:_1 [: u [: v ])"v'
UnderRankVTacitDyadic=: 2 : '([: v^:_1 ([: v [) u [: v ])"v'

-UnderRankVTacitMonadic> 1 2; 3 4
NB. _1 _2; _3 _4
(5;6) -UnderRankVTacitDyadic> 1 2;3 4
NB. 4 3;3 2

NB. &.
-UnderRankVTacitMonadic>
-UnderRankVTacitDyadic>

n=.50
v=.?~n
$v
10{.v

ts'-AtTacit-AtTacit-AtTacit-AtTacit-@- v'
ts'-AtTacit-AtTacit-AtTacit-AtTacit-AtopTacit- v'
ts'-At-At-At-At-Atop- v'
ts'-@:-@:-@:-@:-@- v'

--
For information about J forums see http://www.jsoftware.com/forums.htm