John Hasler <[EMAIL PROTECTED]> writes:
> I don't see why gnucash needs such a thing, though.
We may or may not be interested in it as a database for our financial
data. The current question is simply, does it give us a superior
solution to the "storing things associated with a given account group
that would otherwise have to be in multiple files" problem.
For example, right now we just have the engine data in the .gnc (or
.xac) file, but we need to start storing a lot more data that's
associated with the account group, but not strictly speaking, data
that should be mingled with the engine data, things like budgeting
info (which might actually eventually go in the engine eventually),
recurring transactions, per-user preferences for a given account group
(window placements, colors, whatever), multiple-user locks, etc.
This may or may not be exactly the set of things that we want, but it
seems clear that we need to store more data with each account group
than we are storing now. Before I started thinking about the
Sleepycat DB, the leading contender to solve this problem was the
implementation of a home-grown API that would just store the data in
multiple files in a subdirectory, but this is arguably less intuitive,
and more hassle for the end user, than storing everything in one
"sectional file". This is what the Sleepycat DB (or just the
stripped down version that's in libc) might give us, and it might also
give us a straightforward migration path for the engine data if we
decide we want a fancier storage mechanism.
So instead of having:
my-gnucash-file.gnc/data
my-gnucash-file.gnc/budget
my-gnucash-file.gnc/recurring-transactions
my-gnucash-file.gnc/someuser.lock
we'd just have
my-gnucash-file.gnc
which is a db2 file and would contain the above files as Sleepycat
subname databases.
Anyway, this is the motivation for the current discussion. Hope it
makes my motivation at least somewhat clearer, even if you're not
convinced.
--
Rob Browning <[EMAIL PROTECTED]> PGP=E80E0D04F521A094 532B97F5D64E3930
