On Sun, Aug 4, 2013, at 7:04, Harshad RJ wrote:
> On Sun, Aug 4, 2013 at 1:57 AM, John Wiegley <[email protected]> wrote:
> 
[...]
> These considerations might lead to ambiguities.
> 
> I feel, in the case of ledger, it is possible to take a general stand on
> such problems: "the order of specification doesn't matter". So it wouldn't
> matter if definitions are "made before" or "parsed before" or "processed
> before", etc. The user can expect consistency of results irrespective of
> how the internals of the program behave.

That's exactly why beancount chooses it this way. The file is just a 
database of transactions and entries, assuming that they will later be
sorted by date.



> > Lastly, how do you order three transactions on the same date from three
> > different files, which may even be inter-dependent?
> 
> If we take a stand about consistency as outlined above, then I imagine the
> following solutions:
> 
>    - program throws an error when it finds such cases
>    - user then needs to manually order the specification using numbers or
>    time, etc
>    - alternatively, user opts for "auto-ordering" using a command line
>    flag. This is not on by default.

In practice, I have found it very little of a disturbance to have multiple 
transactions on a same day occasionally show up in a slightly different
order than the actual corresponding "real world" account.  It's not a real 
problem.

There's the occasional (and rate) unexpected intra-day negative balance 
that might show up, but one "fix" is not to render account balance except 
for the last transaction of the day.

-- 

--- 
You received this message because you are subscribed to the Google Groups 
"Ledger" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to