Looking at the J expression to calculate would not make any sense to
someone who does know J. Giving a reference or even putting the original
expression using traditional mathematical syntax and showing their
similarity would show how J is not that hard to read.
On Fri, Feb 21, 2014 at 7:02 PM, Pascal Jasmin <godspiral2...@yahoo.ca>wrote:
> Even if spreadsheets are streets behind (unhip), I think people still want
> to make calculations, and J is both quicker than any other language, and
> much quicker than spreadsheets. A key benefit to powerful one liners is
> interactively obtaining answers.
>
> The way I would enhance the original example would be to enhance it to
> daily interest and more specific payment schedules. It is relatively easy
> concept, that would seem intimidating to code at the same time:
>
> toDailyint =: ^&365 inv
>
> (^&365 inv) 1.05
> 1.00013
>
> looks right, and didn't even need to know the math to find it. other
> language problems.
>
> 1.000131 ^ 365
> 1.04859
>
> could point out the extended precision option x:@(^&365 inv) if there is
> a need to get it right, but the above is probably the number the
> bank/lender would use.
>
> Lets say you want to make payments every 2 weeks.
>
> 28$ 1,~ 13 # 0
> 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1
>
> shows payments for 28 days.
>
> 3650 $ 1,~ 13 # 0 NB. payment days for 10 years.
>
> 0 0 0 , 3650 $ 1,~ 13 # 0 NB. if payments start 2 weeks from 3 days from
> now. (then 2 weeks thereafter)
>
>
> biweekly =. 40 * 3650 $ 1,~ 13 # 0 NB. $40 payments biweekly for 10
> years.
>
>
> ] endofyearmarks =: <: 365 * >: i.10 NB. will be used for
> display/classification and bonus payments.
> 364 729 1094 1459 1824 2189 2554 2919 3284 3649 NB. <: used to compensate
> for 0 start.
>
>
> yearly =. 100 endofyearmarks} 3650 $ 0 NB. pay $100 at end of each year.
>
> 100 (endofyearmarks - 30)} 3650 $ 0 NB. would set payment date 30 days
> prior to end of each year.
>
>
> payments =. biweekly + yearly NB. add lists of payments like a
> spreadsheet would. Reason to calculate separately and sum is that payments
> might occur on same days. Cool feature of J, that drew me into the
> language. When I was learning it, I called these "noun collisions"...
> smashing together data structures into 1.
>
> (0,endofyearmarks) { ([-~(toDailyint 1.05)*])/\. 10000,~|.payments
> 1630.09 2663.44 3647.23 4584.05 5476.11 6325.57 7134.43 7904.64 8638.04
> 9336.39 10001.3
>
> One issue with the above, is that payments is a compound noun without any
> major benefits to it being so. The advantage of single powerful lines is
> that we don't have to recompute nouns for instance if we modified biweekly
> or yearly, or added a new payment stream. To balance that and readability,
> a better version of the statement is:
>
> (0,endofyearmarks) { ([-~(toDailyint 1.05)*])/\. 10000,~|. biweekly +
> yearly
> 1630.09 2663.44 3647.23 4584.05 5476.11 6325.57 7134.43 7904.64 8638.04
> 9336.39 10001.3
>
> Darn.... I was hoping to have the loan paid off in 10 years.
> There needs to be a modification to find the day that a 0 balance is
> reached, but a cool reshaping feature is available
>
> (0,endofyearmarks) { ([-~(toDailyint 1.05)*])/\. 10000,~|. (12*365)
> $ biweekly + yearly
> _590.933 546.808 1630.31 2663.44 3647.23 4584.05 5476.11 6325.57 7134.43
> 7904.64 8638.04
>
> shows results at the end of years 2 to 12, and tells us that full
> repayment is made sometime in year 11. To find exact date:
>
> (1 i:~ 0>:]) ([-~(toDailyint 1.05)*])/\. 10000,~|. (12*365) $ biweekly
> + yearly
> 170
>
> 170 days before end of year 12.
>
> To me, there would be major struggles in accomplishing this in another
> language, and a spreadsheet probably doesn't have enough rows.
>
> A good punch line is to show J performance by extending the example to a
> $23100 loan paid over 100 years... by simple modification to edit history
>
> (0,endofyearmarks) { ([-~(toDailyint 1.05)*])/\. 23100,~|. (100*365) $
> biweekly + yearly
> _4164.15 _2855.63 _1609.02 _421.909 708.536 1785.02 2810.1 3786.23 4715.75
> 5600.86 6443.7
>
> I found 23100 by trial and error, by adjusting it until the last 10 years
> crossed the 0 point, but it shows off J's speed in calculating through
> 40000+ rows instantly, and everything else that happens on that line.
> Never mind the pain of copy pasting 40000 rows in a spreadsheet, updating
> the starting loan value would bring it to its knees.
>
>
>
> ----- Original Message -----
> From: Brian Schott <schott.br...@gmail.com>
> To: Programming forum <programm...@jsoftware.com>
> Cc:
> Sent: Friday, February 21, 2014 6:51:02 PM
> Subject: Re: [Jprogramming] J in 5 minutes
>
> Don,
>
> My intent was not to make the essay a financial tutorial, but to highlight
> the beauty and uniqueness of J code, so some of those details seem
> inappropriate. I need another introductory sentence/statement or else I
> need to change the example.
>
> Thanks for you help,
>
>
>
> On Fri, Feb 21, 2014 at 4:51 PM, Don Guinn <dongu...@gmail.com> wrote:
>
> > I'm confused! I assume that the annual payment against the $10000 loan is
> > $1000. Which means that the value of the loan must be more than $7000 as
> > only $3000 has been paid against it. Yet the line which should be the
> value
> > of the loan after three years shows $3711.05.
> >
> > I think the description of the problem needs to include the amount of
> > payment and if the payments are annual or whatever. Also, it would help
> to
> > reference the formula used, such as a link to a web page deriving that
> > formula rather than seeing it in J without an explanation.
> >
> > And I couldn't view the video either.
> >
> >
> >
> --
> (B=)
> ----------------------------------------------------------------------
> 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
