gnucash2 C binary, cashutil and scheme

[Chris Shoemaker]

On Wed, Jan 11, 2006 at 11:28:49PM +0000, Neil Williams wrote:
> On Wednesday 11 January 2006 11:05 pm, Derek Atkins wrote:
> > However, I do want to ask how this distraction is going to help get
> > G2 out the door?  And yes, I do consider this is a distraction right now.
> 
> Well, if we want a C binary to replace the scripts for G2, then it's fairly 
> complete. From your reaction, I assume we don't now? I thought we did last 
> week. It's not dependent on cashutil, it could be implemented in trunk 
> without all the other cashutil build stuff. It just uses some code that I 
> developed for cashutil/pilot-qof. (i.e. the CLI component works).

Derek's question _could_ be interpreted as discouraging any change to
Gnucash that's not directly and obviously related to releasing G2.
Let me offer a slightly different perspective.  I don't actually know
what changes are in that theoretical set, so here's what I do: When I
consider making some particular changes, first I ask if the changes
are technically valuable.  Then I ask if they might _delay_ or
_interfere_ with any potential Gnucash release plans.  

My experience has been that nobody complains at development that's
both technically beneficial, non-distracting and non-interfering, even
if it's not clearly preparing a release.  So, my advice (for any
gnucash developer) is: don't avoid making changes just because you
don't think they're release-related.

OTOH, large atomic commits that _almost_ work, followed by requests
for help in integrating them _are_ a bit distracting, even when
they're benefical changes, so I think Derek's concern is valid.

-chris