General Ledger/Accounting Libraries

[Frank Millman]

Andreas Pauley wrote:
> Hi All,
> 
> Our company is currently evaluating Python as a language for writing
> financial/accounting type software (among others).
> 
> What libraries or packages are available in this domain for use?

Hi Andreas

I was in two minds as to whether I should respond to this. I am
working on something that may provide what you are looking for, but it
is still under development. I don't want to waste your time with
vapourware, but if I explain what my ideas are and how far I have got,
you can decide whether to take it any further.

I am developing a generalised accounting/business package using
Python. For the GUI I use wxPython. For the database, I use PostgreSQL
with pyPgSQL on Unix/Linux, and MS SQL Server with Win32/odbc on
Windows.

It is too big to describe in a few lines, but I will list some of my
design goals, and discuss how close I am to achieving them.

Data integrity -

I am an accountant first and a programmer second, so I have this area
well covered. Some programmers are reliant upon input from an
accountant/business analyst, and problems can arise when there is a
communication gap. I wear both hats, so I just *know* what is
required. In 20+ years of writing accounting software, my systems have
received a number of criticisms, but going out of balance is not one
of them.

Performance -

Using a database cursor and a wxPython virtual grid, I can select and
display 10 000 rows out of a table of 260 000 rows in less than a
second. I can scroll up and down the table with no perceptible delay.
I can look for a particular row based on a search string, and it finds
the row almost instantaneously. This is on a 266mhz Pentium II laptop
with 128mb ram and Windows 2000. On my linux box with a 1.7ghz
Celeron, it is even faster. I am reasonably confident that I can
handle large volumes of data without a problem.

Access control -

Full access control is built in from the ground up. Users must log in,
and are authenticated. Each user is assigned to one or more groups,
and each group can have read/write privileges defined on every column
in every table in the database. Full audit trails will be available
(not done yet, but on the to-do list). Obviously I cannot prevent
unauthorised access external to my software, so it is up to the system
adminstrator to control access to the database itself and to the
python programs.

Flexibility -

I used to be a fan of customisation, but with hindsight I believe that
in the long run the benefits are illusory. I am therefore building in
a lot of flexibility in such things as structure of g/l code and
product code, ability to add your own business logic to database
tables and columns, parameters to select things such as ageing
methods, etc. Full source code will be available, so anyone can
customise the programs if they want to, but I will be very open to
requests to fold such customisations back into the main body of code,
possibly selectable by a parameter.

Support -

The source code will be released under an Open Source license. Support
will be available at a modest monthly charge (to be determined).
Although I will have a few sites that I will support directly, my
intention/hope is to develop a network of dealers/consultants, who
will install and maintain the software at end-user sites and provide
first-line support. I in turn will provide backup and support to them,
in addition to continuing ongoing development of the software.

Current state of software -

I am getting close to setting up my first few live installations - say
1 to 2 months. The initial release will include General Ledger, Sales,
Purchasing, Inventory, Accounts Receivable, and Accounts Payable. It
will be multi-company, multi-branch, and multi-currency.

Where to from here -

It all depends upon the success of my business model (see Support
above). If I can get a few dealers to get a few installations, and
start to generate a revenue stream from support fees, I will plough
this back into providing a full support and development
infrastructure, employing staff as needed. My wish-list for ongoing
development of the software is endless, but includes things such as
full context-sensitive help, multi-language support, image support,
web front end, etc. There will also no doubt be many requests for
additional functionality. This can turn into a virtuous circle - the
more I build in, the more people will want to use it, which will
increase the revenue stream, which I can plough back into development,
etc.

However, I have to be realistic, which means I must get this past the
vapourware stage and get a few live installations. My inclination is
to not even release the source code until this has settled down, as
there are bound to be teething problems. I estimate that this will
take up to 3 months. I will me more than happy to discuss
requirements, concepts etc in the meantime.

Any comments will be very welcome.

Frank Millman