No-power version:
addx =. , @ (2&({.!.'x');.1~ >:/\.&.|. @ (1 , }. = }:))
addx 'allisdiscoveredflee'
allisdiscoveredfleex
addx 'callisdiscoveredflee'
calxlisdiscoveredfleex
addx 'xallisdiscoveredxflee'
xalxlisdiscoveredxflexex
Henry Rich
On 2/20/2011 11:12 AM, Marshall Lochbaum wrote:
> ({.={:) works the best: not only does it give the desired functionality, but
> it inserts a fill at the end for odd numbers of elements.
>
> preprocess=: (#!.'x'~ 1 j. i.@# = (2* 1 i.~ _2 ({.={:)\ ])) ::]^:_
>
> preprocess 'allisdiscoveredflee'
> allisdiscoveredfleex
> preprocess 'xallisdiscoveredflee'
> xalxlisdiscoveredfleex
> preprocess 'xallisdiscoveredxflee'
> xalxlisdiscoveredxflexex
>
> Even running
> preprocess 'x'
> will go forever with this. The Playfair cipher has no mechanism for dealing
> with double x's or ending x's. Best is just to not put any x's in your
> initial string.
>
> Also, the preprocessor should replace all j's with i's:
>
> replace_j_i=: (=&'j')`(,:&'i')}
>
> (,: replace_j_i) (97+i.26){a.
> abcdefghijklmnopqrstuvwxyz
> abcdefghiiklmnopqrstuvwxyz
>
> Marshall
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Henry Rich
> Sent: Sunday, February 20, 2011 9:18 AM
> To: Programming forum
> Subject: Re: [Jprogramming] Playfair cipher
>
> (=/`0:@.(2|#))
>
> could be =/@:(2&{.)
>
> but if it doesn't matter what the result is when there is one character
> (as it appears to my eye), ({.=}.) would also work. Best might be to
> use fill: =/"1@:(_2&(]\))
>
> Insert-at-index is a pain, and I don't know of any better way than using #.
>
> I spent some time thinking of a way to do this without ^:_ with no luck.
> I'd like to see one.
>
> I'm afraid to run preprocess on an input of 'xxx': it looks like a slow
> way to out-of-memory.
>
> Henry Rich
>
> On 2/19/2011 9:16 PM, Marshall Lochbaum wrote:
>> dblprocess=: (#!.'x'~ 1 j. i.@# = (2* 1 i.~ (_2 (=/`0:@.(2|#))\ ])))
>> ::]^:_
>> preprocess=: [: ,&'x'^:(1=2|#) dblprocess
>>
>> Provides a moderately blunt preprocessor. Things I'd like fixed:
>> (_2 (=/`0:@.(2|#))\ ])) should just be (_2 =/\ ]) , but on an odd length
>> string the last part, say (=/\ 'e') , gives a single character and not a
>> Boolean, resulting in a domain error.
>> Does anyone know a better way to do insertion with an index than using # ?
>>
>> preprocess 'allisdiscoveredflee'
>> allisdiscoveredfleex
>> preprocess 'xallisdiscoveredflee'
>> xalxlisdiscoveredfleex
>> preprocess 'xallisdiscoveredxflee'
>> xalxlisdiscoveredxflexex
>>
>> Marshall
>>
>> -----Original Message-----
>> From: [email protected]
>> [mailto:[email protected]] On Behalf Of Henry Rich
>> Sent: Saturday, February 19, 2011 1:40 PM
>> To: Programming forum
>> Subject: Re: [Jprogramming] Playfair cipher
>>
>> Oh yeah,
>>
>> pf =. 'paswo','rdbce','fghik','lmnqt',:'uvxyz'
>>
>> (thanks Ric)
>>
>> Henry Rich
>>
>> On 2/19/2011 11:39 AM, Henry Rich wrote:
>>> I worked on the easy part:
>>>
>>> pfr =. ((] (<"1@[ { ])~ ((0 1&(|."_1)&.|:@[)`(-"1 |.)@.(+./@])
>>> =/) @ ($@] #: (i.~ ,))) ];.0)~
>>>
>>> ; _2 pf&pfr\ 'allisdiscoveredfleex'
>>> pmqfabhwewzddrrgtrbz
>>>
>>> ; _2 (];.0 pf)&pfr\ 'pmqfabhwewzddrrgtrbz'
>>> allisdiscoveredfleex
>>>
>>> I can now add monad u;.0 to my Life List.
>>>
>>> (you would also need a preconditioner to handle double letters and
>>> odd-length input).
>>>
>>> I was hoping the same program would work for the hard part, but I
>>> wasn't sure what to do when the two letters match in more than one axis.
>>>
>>> Henry Rich
>>>
>>>
>>> On 2/19/2011 7:40 AM, Kip Murray wrote:
>>>> The Playfair cipher http://en.wikipedia.org/wiki/Playfair_cipher is
>>>> based on a 5 by 5 character array that treats i and j as equivalent:
>>>>
>>>> paswo
>>>> rdbce
>>>> fghik
>>>> lmnqt
>>>> uvxyz
>>>>
>>>> To encode "All is discovered flee" you replace pairs of letters al li
>>>> sd is co ve re df le ex with pairs diagonally opposite, one to the
>>>> right, or one below (wrapping) from the table, getting
>>>>
>>>> pm qf ab hw ew zd dr rg tr bz
>>>>
>>>> Your mission, should you decide to accept it, is to write j code for
>>>> enciphering and deciphering, given the five by five table. That's
>>>> the easy part. The hard part is to design and implement an analogous
>>>> cipher that uses the 3 by 3 by 3 character array below to replace
>>>> triples of letters alx _is _di sco ver ed_ fle ex_
>>>>
>>>> pas
>>>> wor
>>>> dbc
>>>>
>>>> efg
>>>> hij
>>>> klm
>>>>
>>>> nqt
>>>> uvx
>>>> yz_
>>>>
>>>> ---------------------------------------------------------------------
>>>> - 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
>>
> ----------------------------------------------------------------------
> 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
