Storage API

Most accounting systems continually reimplement double-entry accounting. As a developer, I find this frustrating because very commonly, user-interface code for accounting systems is co-mingled with details of the double entry accounting implementation.

What I propose is a clear API that simply does the basic functions of double-entry accounting, treating double-entry accounting more like mathematical operations and less like business logic. Business logic varies, but the rules of double entry accounting remain roughly the same.

Users of the API would be those who write accounting applications and want to treat the double-entry portion purely as a black box.

The ideal scenario would be an known double-entry accounting API that different accounting projects could support, separating the problem of storage of double-entry accounting data from specific accounting systems.

A very rough idea of the API's core data structures follows. This is not fully baked.

Postings

The primary record of a double-entry accounting is a posting, which has the following fields:

Each posting has a globally unique identifier that does not change after its first creation.

State

State has the following fields:

The idea behind posting state is that it is somewhat user-configurable. Typical posting states would be things like CLEARED for items that have cleared a bank account. However, one idea behind the state is use it to allow for "user-specific" versions of the books. The goal is to allow users to "stage" postings that may or may not be accepted. For example, a user might submit an expense report, and the state might be "REQUESTED". Generally, the rest of the system would ignore REQUESTED postings, as they aren't officially approved for the books yet.

Entry

Entries have the following fields:

Each entry has a globally unique identifier that does not change after its first creation.

Storage in RDF

It was suggested on the mailing list that RDF might be an interesting format to explore for representation of the stored data. This idea might be worth exploring, but the open question is whether or not the StorageAPI needs to specify a specific format?

Regarding this, review of Up The Asset and 3account might be useful.