year end, already !

[Derek Atkins]

Andrew Sackville-West <andrew at farwestbilliards.com> writes:

> Derek, how does that aspect work anyway, as far as spec'ing what parts 
> of the app should do? is it a seat of the pants thing? committee? user 
> input? just curious.

It depends on the size of the feature.  For example I spent a couple
months writing up design docs and running them by the -devel list
before I started in on the Business features.  But admittedly that's a
relatively LARGE feature.  Something small can usually just be
proposed and then implemented.  Sometimes it's just implemented and
sent in blindly, or a current dev will "just do it".

In general we like to at least talk about things ahead of time.  While
we don't have a written, complete architecture document for gnucash,
we do have some 'lore' that's usually useful to help new devs or to
bring features along.

On the other hand, I must admit that "running code" has a lot going for
it.  :)   If you can submit a working patch it's more likely to get your
feature into the code than if you just submit an RFE in bugzilla.
(It's MUCH easier to audit, apply, and test a patch than it is to implement
a feature from scratch).

Hopefully this rambling did actually answer your question. :)

> SURELY
>> some bright light at some big company will see that the project is great
>> but needs some husbanding.
>> Might we mount an "adopt-an-application" campaign ? Y'know; say it with
>> money, enough of it, and on-going !
>> Who wants to be the forward thinking company behind gnuCash ?
>
>
> Hear hear! Or how about another tack -- I own a small business. I would 
> happily put up X dollars a year to pick a function and pay for it to be 
> finished up. If we can gather a few like minded people together, each of 
> us kick down so many dollars and then the developers can put in the time 
> and finish up one aspect at a time?  Maybe there could be an estimated 
> cost per feature list and then small groups of people could layout the 
> dough for features they want to see implemented? Example -- developers 
> can say  "we think it takes 20 hours to implement X" at $y per hour that 
> costs $(20 x y). Start a campaign to raise that money and when its 
> raised, it gets done?

This idea has been suggested before.  The problem is sort of
chicken-and-egg.  It's hard to gather enough momentum to get the
process started.  And then it's hard to dole out the cash once you
have it.

> Just brainstorming...
>
> Andrew

-derek

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available