Re: GnuCash needs developers (esp for GTK/GNOME/Guile work).

[abel@bfr.co.il (Alexander L. Belikoff)]

Rob Browning <rlb@cs.utexas.edu> writes:

> GnuCash is a program designed to manage personal finances (a la
> Quicken).  We either already support, or have plans to support
> features like normal accounts, investment accounts, on-line banking,
> automatic brokerage account (stock) updates, etc.  Currently GnuCash
> needs more good developers.
> 
> Right now, the stable interface is Motif/Lesstif, but we are in the
> process of creating a GTK/GNOME interface, and we need people
> interested in working on this.  Most critical at the moment is the
> development of a reasonable GNOME/GTK register.  What we have now is a
> total hack built around a CList.
> 
> Also, we've recently integrated Guile, and now we need to put it to
> work.  For example, we want to add a Emacs-like "hook" system, an
> engine for handling key-bindings, etc.  I haven't looked carefully
> yet, but the Emacs source is probably a good source of ideas if not
> actual code.
> 

I'm currently collecting resources on personal finance management,
since I got to a point where I need a personal finance program. And I
want it to be GPLed and run on UNIX.

A couple weeks ago I took a look at various programs available. Most
of them are not useful at all - they allow just simple things. Looks
like all developers write programs beginning with a GUI part, leaving
the real functionality for later. Pretty pathetic...

My search resulted in an idea of 'yet another' PFM. The idea was (and
still is) to write a program in Guile. Guile (Scheme) is perfect for
prototyping and rapid development, and we are not talking about an
application requiring time-critical computation. The GUI would be GTK,
but it'd be coded only after the 'meat' works.

I don't set any ambitious goals, like automatic tax management, but
apart from that, the program must do *real* things - it shouldn't look
like a toy compared to MYM or Quicken. It may not have as elaborate
GUI, but it will have most of their functionality. At a later stage,
I'll add some support to automated banking (if it is possible to
implement without paying huge royalties to the technology godfathers).

There are two things preventing my coding right now. First is my
complete ignorance in personal finance management. This will be
changed - I'm reading some stuff, so I'll be set at some stage.

The second reason is much more serious. I don't have a carte blanche
for writing freeware programs. In general, Bloomberg doesn't allow
employees to write any freeware at all. I was an exception to get a
list of 'approved' programs, I am allowed to write under GPL terms.
And it doesn't include a PFM. I'll negotiate about it at some stage,
but, considering Bloomberg primary business, they may be less than
enthusiastic to allow any GPLed program, containing a word 'finance'
in its description. And I definitely won't be allowed to include a
mortgage calculator in it. :-)

Anyway, if GnuCash people are open-minded enough, we could start
working on a spec in the near future. Yes we need an extensive
specification of *what* we are writing - not just vague ideas. Once
this is done, we would start coding the engine in Guile. Once it
works, a GUI time will come.

Regards,

-- 
Alexander L. Belikoff
Bloomberg L.P. / BFM Financial Research Ltd.
abel@bfr.co.il