Hi Ric,
You wrote:
> ''Let
> mean=: (+/y) % #y
> ssdev=: +/ *: y - mean y
> var=: (ssdev y) % <:#y
> stddev=: %: var y
> )
Tacit J is a central component of J that is introduced early in the
documentation All I really want is a mode in which I can remove the
surrounding material and just write:
mean=: (+/y) % #y
ssdev=: +/ *: y - mean y
var=: (ssdev y) % <:#y
stddev=: %: var y
stddev2=: %: (% <:#y) * +/ *: y - (+/y) % #y
That's not a lot to ask. I'll give up on replacing the x and y with [ and ].
Anyway, I have some work to do, I'll be back.
Don
----- Original Message -----
From: "Sherlock, Ric" <[email protected]>
To: "Chat forum" <[email protected]>
Sent: Friday, May 01, 2009 3:25 AM
Subject: Re: [Jchat] Language S
> Don,
>
> So far the all the evidence and advice brought up in this thread suggests
> that:
> * explicit reference to arguments in equations is likely to be less
> confusing to math students (or beginning J users) than tacit or
> implicit references.
> * the use of a little additional syntax to define verbs presents
> little or no barrier to students unfamiliar with J and/or
> programming.
>
> I wonder at this point whether you would be receptive to going back to an
> earlier suggestion in post by Oleg proposing a "verb definition mode"?
>
> The implementation details are specified on the wiki page:
> <http://www.jsoftware.com/jwiki/Phrases/Definitions>
>
> If the adverb "defn" on that page were renamed "Let" you could use it in a
> session like this:
>
> ==============
> NB. blah blah intro
> NB. For:
> y=: 3 4 9 4 3 8 5
> NB. The mean is
> (+/y) % #y
> 4.6
>
> NB. ... when want to define verbs:
>
> ''Let
> mean=: (+/y) % #y
> ssdev=: +/ *: y - mean y
> var=: (ssdev y) % <:#y
> stddev=: %: var y
> )
> mean y
> 4.6
> ssdev y
> 29.2
> var y
> 7.3
> stddev y
> 2.70185
>
> ==================
>
> If you wanted to get really picky you could define a noun
> Now=: 'whatever you like cause it just gets ignored'
>
> So that you could write:
>
> Now Let
> coeffvar=: (stddev y) % mean y
> )
>
> That's getting pretty close to the treatment of equations in a text book?
>
> Note you are free to make the definition for stddev one big line if you
> like, I just think that dividing it up into meaningful bits as above is
> more illuminating (not to mention flexible).
>
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
