Moving this to chat as this seems quite a way from programming.

I prefer to use p. as well. But when I saw the original forum question it
got me to remembering trying to do proofs tacitly.
I should have tested the statements using J. But one should be able to do
this without such a crutch. I struggle doing proofs using tacit expressions
instead of explicit expressions like I did in high school algebra and trig.
It seems that somehow it ought to be as easy but for some reason I make
serious mistakes like this one.

Then there is the thought that intrigues me as I try. There's always the
approach to write a verb to test each line of the proof with various
arguments which would have caught this error. There might even be a way to
write a verb which will verify the correctness of each step. And maybe even
one that would generate the steps get from one statement to another thus
generating a proof or show that a conjecture is false. In doing proofs I
used various proven identities to transform one statement to another. We
have such identities implied in the J dictionary and from mathematics like
associative, commutative and distributive. Applying them seemed as a
mechanical process but there seemed to be a bit of art in choosing
identities to use.

So far I'm stumped.

On Wed, Aug 19, 2009 at 8:17 AM, Raul Miller <[email protected]> wrote:

> On Wed, Aug 19, 2009 at 10:12 AM, Zsbán Ambrus<[email protected]> wrote:
> > ([^2:)+(2**)+(]^2:)
> > ((2^~[)+(2**)+(2^~]))
> > (^&2...@[+(2**)+^&2...@])
> > (*:@[+(2**)+*:@])
>
> Note that these are not equivalent (the first
> one gives a different result than the later
> examples).
>
> (When I am simplifying tacit expressions, I
> always like to include a representative
> argument so I can have the machine catch
> simple mistakes.)
>
> That said, personally, I like using p. for polynomials.
>
> FYI,
>
> --
> Raul
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to