Raul,
Here are some examples from my version. I no longer have access to SAS.
Granted that 0.1*i:12 takes some preprocessing.
pretty SL >.0.1*i:12
_0 | 11100000
0 | 00000111111111122
pretty 2 SL >.0.1*i:12
_0 | 11100000
0 | 000001111111111
2 | 22
pretty 5 SL i: 12
_10 | 21
_5 | 98760
_0 | 54321
0 | 01234
5 | 56789
10 | 012
On Thu, Mar 14, 2013 at 4:42 PM, Raul Miller <[email protected]> wrote:
> On Thu, Mar 14, 2013 at 4:23 PM, Brian Schott <[email protected]> wrote:
>> I mentioned SAS before. SAS allows the use of stems that are either
>> X%10 5 2, I believe. I believe my version also allows such.
>
> Some examples would be good, to resolve the questions this raises for
> me. i:12 and 0.1*i:12 are probably good values to use in the
> examples.
>
> It might be that Kip's most recent example, with stems like t, f, s,
> e, 1 is an example of the X%2 case. But if that's the case, note that
> X%2 is not the stem here. 10%2 is 5.
>
> That said, if the requirement is that the algorithm implicitly pick
> the stem algorithm based on the data, and if the requirement is to
> exactly duplicate SAS, and if there's no human understandable complete
> definition that we can use, what we have is a requirement to use SAS.
>
> --
> Raul
> ----------------------------------------------------------------------
--
(B=) <-----my sig
Brian Schott
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
