On Sat, 24 Jun 2000 21:42:22 EST, the world broke into rejoicing as
Richard Wackerbarth <[EMAIL PROTECTED]> said:
> On Sat, 24 Jun 2000, Robert Graham Merkel wrote:
>
> > Oh, and one other thing. Performance when doing monetary arithmetic
> > is currently not an issue. As far as I can see, this isn't likely to
> > change, whatever implementation we end up using. Is this the case?
>
> I think you are wrong. When you have a large database of transactions,
> the "ripple effect" of a single change can cause a huge number of
> balances to be recomputed. Performance can be a real concern.
But must be balanced _somewhat_.
The "ripple" effect can head in two directions:
- When the "huge horde of balances" pop up to be recomputed, it may
very well be that this indicates that something is wrong. There
should not be thousands of things getting recomputed.
- The bottleneck may _still_ not be with monetary arithmetic.
If the GUI has to generate a thousand lines worth of register
items, it won't be the monetary items that are the problem, but
rather setting up the register widget.
The real point here is that so long as we _consider_ efficiency,
it does not necessarily have to be the _only_ consideration.
If we wind up with numeric operations being only half as efficient,
it's hardly relevant to overall performance if it's only increasing
from 3% to 6% of the computations...
--
[EMAIL PROTECTED] - <http://www.ntlug.org/~cbbrowne/linux.html>
Rules of the Evil Overlord #151. "I will not set myself up as a
god. That perilous position is reserved for my trusted lieutenant."
<http://www.eviloverlord.com/>
--
Gnucash Developer's List
To unsubscribe send empty email to: [EMAIL PROTECTED]
