On 04/02/2018 13:44, Christopher Lam wrote:

I wished to experiment in what budgeting should look like by using the existing engine, UI, and reporting infrastructure.

If you want to know how one form of budgeting, that which includes envelope budgeting budgeting is likely to be implemented from a working point of view I have already said look at ledger-cli, beancount and friends. gnc is *not* going to reinvent the excellent work that has been done there for the simple reason that it would be a waste of resources.

The UI is, of course, the problem hampering gnc vs keeping up with the family.

It's actually not that difficult to create a 'budget balance calculator'; whether it meets the needs for everyone is another matter. But for people who wish to experiment with envelope budgeting, I can confirm that it is possible.

I know that! There are lots of free spreadsheet versions. The bit that baffles me is why some people pay someone else to say budget this way rather than that.

Rules are:

  * budget transactions must be "outside the books" so to speak, i.e.
    they are not counted in any net worth, profit/loss, transaction
    report, etc. they must be, by default, invisible to the reporting

yup, you read from the db for that

  * in order to be acceptable by the engine, they should they should
    satisfy the double-entry equation.

nope, informal budgets can include single entry, they can include "inheritance from Aunt Mabel if she dies before Christmas"

  * it should be generally useful

there lies the rub, generally useful to which group of users ?

  * it should be better than the current budget

or less bad :)

So this is actually a success.

Did you vote for Trump?  Success seems an obscure term these days.

I don't use transaction voiding, and have hacked "voided transactions" to become "budget transactions" and upgraded the status bar display. The results are available on the topmost commit in https://github.com/christopherlam/gnucash/commits/envelope-budgeting as a demo of "what can be achieved using the engine". But I stress this was an experiment.


You have made a basic error in that the budget impinges on actual.

Think of it from a business POV, "We'll buy 100 widgets next month if the price is right", your budget, if I am reading it right, doesn't see that as a plan but an action. No way will that ever hit real code.

Also voided tx although little used, actually do something if only to stop other things happening at times. I think they might already be re-purposed in a way.


Christopher, have you got an sql backend working ? Ideally with Sebastian's python bits as well ?


gnucash-devel mailing list

Reply via email to