On 10 April 2011 23:48, Viktor Cerovski <[email protected]> wrote:
>> To start off, the Dictionary entry only says that identities like
>> the cited one are extended to certain cases.  It does not say
>> whether that particular identity holds.  And of course, for most
>> values of u and y, the identity does *not* hold.  It is not a
>> general identity.
>>
> I agree.

Then we should also agree that we cannot use the said identitiy
to draw conclusions about what u/y does when 1=#y -- as you
initially stated.

>> For the particular case of one-item y, witness:
>> ..................
> Your complaint here is presumably that the last two values
> are not equal.  My answer to this, in short, is that they
> cannot be equal in the case of Minus.  Here is why:
> ....

Of course they cannot, and I did not expect them to be.  I was
just showing an example why the `identity' is not actually one,
except in very special circumstances, and that therefore you cannot
derive u/y's result at 1=#y from that `identity'.

More generally, the `identity' would not hold for functions that
are not (algebraically) associative or do not possess a neutral value.

Even for associative functions with neutral values, it will still not hold
when 1=#y and y is of inappropriate type for u -- the r.h.s. then is an
incorrect expression, while the l.h.s. (u/y) apparently is assumed correct.

> Well yeah, sure, the "identity" does not hold because each side
> of the identity cannot be evaluated.

Actually, the r.h.s. fails, while the l.h.s. doesn't -- which, as I said,
once more renders the identity useless as a means to draw conclusions
about / in general.

> ..................................
> I hope the above clarifies it a bit.

The problem is not that I do not understand why and how the
`identity' is broken.  I do understand that very well (I am a
mathematician by education).  The problem that I see is with
the definition and the actual operation of / -- the former
appears to be insufficient w.r.t. details, the latter is
complicated and has anomalies.  That's all -- just an observation
-- I don't really suffer because of it  :)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to