Re: Simple Interest Calc Design

Rob,
For the current interest calc system i'm working on, we post interest income once per month or statement period depending on the users'/system setup choice. Interest per day is computed on the daily balance, but posted as a dr to the customer's account in one lump per period. When a transaction is posted to a prior period, the entire month or period is recalculated and the incremental amount of interest adjustment is also posted.

For a previous credit system, there was a requirement to track each debit's corresponding credit. For this, we kept a FK to the related check, PO, or invoice in the transaction record.

A final system I only reviewed for insight and to learn lessons, used a balance forward method. This entailed adding up every open transaction, summing them, marking them as closed, and entering a new record in the same table called "bal fwd" that contained the sum only. This seemed weak in that you had calculated data written to the table.

In article <37274f92_at_discussions>, "Rob K" <rfk_at_credaccept.com> wrote:
>
>I was wondering how some companies design
>their system to handle simple interest
>calculations. The system I am working on
>stores not only the debits and credits for
>each account but also what debits are paid
>by which credits. Calculating the simple
>interest is straightforward once you sum up
>the principle balance. The problem arises
>when a correction to a debit or credit back
>in time. Obviously this correction could
>have an effect on the principle balance of
>the account and any interest charges there
>after would be wrong. The only way I see of
>bring the account in line would be to
>recalculate the principle and interest
>balance disregarding any corrections and
>compare it to the current balance, any
>differences would have to go in as
>adjustment transactions. It even gets more
>complicated if a credit needs to be applied
>back in time. i.e. reapply a check that came
>in 3 months ago from one account to another
>account.
>Any comments would be appreciated.
>Thanks Rob
>
>
>
>--Posted from EarthWeb Discussions. http://discussions.earthweb.com
>
Received on Sat May 01 1999 - 03:03:32 CDT