On 10/02/2019 12:40, David Carlson wrote:
Lets all step back a second and consider what the report is supposed to
show before deciding whether it is always correct.
I think the report is conceptually broken.
The report is supposed to cover all of the assets and liabilities that were
selected when configured only over time interval which was also selected
during the configuration.
It doesn't do that as far as I can tell.
It is supposed to use a method of valuation for
accounts not in the base currency which is also selected during the report
configuration.
It doesn't do that, actually.
My reading of the code (and I am an acknowledged fool when it comes to
scheme) is that it uses a balance sheet type approach rather than a p+l
type approach to time. If I am right then the report is wrong to begin
with, because a trial balance can be for either,
a P+L type report (i.e. tx from begin date to end date)
OR for
a Balance Sheet type report (i.e. start at the beginning up to a point,
what have we got?)
Over the last several years there have been a number of discussions about
whether the various valuation methods work correctly, with respect to
report dates vs dates available in the data file and even whether the
values should be displayed or calculated with exact i.e. fractional numbers
or in floating point. All of that is in the infrastructure, not in the
actual report code.
That is incorrect. A trial balance is a report, it shouldn't be doing
valuations at all *except* right at the end, maybe. If it is going to
do valuations they should be presented separately from the account balances.
Remember flox, at trial balance (and it can be from any date to any
date) you have something like
100 in your own currency
1 in your fantasy currency
10 shares (or similar) in investment A
11 units (or similar) in investment B
etc
*that* is what a trial balance should report, the *money* value of
investment A, Investment B and the fantasy currency is what the Balance
Sheet, etc is for, the TB is a different animal.
I believe there were some code changes implemented over the life of the 2.6
code series, and there were more changes in the transition to 3.x.
I don't think we're getting anywhere (I have read the bug thread)
I think the bottom line is that yes, different code releases will give
different results with some data even if the report configurations are
scrupulously checked to be identical.
That's scary.
Now, where do we want to go from here?
I think we need a new TB. The problem is it can't be implemented
however good it is because we're stuck with Scheme reports.
--
Wm
_______________________________________________
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel
