Re: Allow donations for precise bugs

[Sean Middleditch <elanthis awesomeplay com>]

On Sat, 2005-03-12 at 09:12 +0100, Maurizio Colucci wrote:
> Dear friends,
> 
> I apologize in advance for the cross-post but it seems adequate here, as this 
> proposal involves users and developers.
> 
> 
> IN BRIEF
> --------
> 
> Currently, in bugzilla, I can file bugs or propose new features. My idea is 
> to allow users to donate money directed towards a *precise* feature or a 
> *precise* bug.

This is a very bad idea for one primary reason.  Users often request
insanely stupid features.  This isn't a "the user doesn't have a
problem" - it's, "the user invents a solution to their problem and
specifically requests that single solution, instead of identifying their
problem to the developers and letting the developer find a more optimal
solution."

I deal with this all the time in my own projects.  Users come in asking
for some feature that does something the software couldn't do before.
I've made it a habit to ask why they need the software to do this.
After a little discussion, I get to the root of what the user wants, and
can in many cases offer a radically different solution/feature than what
the user asked for, but which still solves the user's true problem and
in a way that both I and the user agree is far better.

Sometimes, though, the user can't distinguish between the real problem
and the solution they propose.  They get it stuck in their head that
there is only one way to accomplish their goal and refuse to even
*consider* an alternative.  If developers were forced to listen to that
user's demands, they'd be forced to implement a suboptimal solution to a
problem that results in an over-all poorer software quality.  The user
might think they're better of because they have this new feature, but
that user is now stuck with something that doesn't work as well as the
developer's alternative, and all other users - even if they see the
benefits of the alternative solution - are pretty much stuck with the
one user's demands, unless you massively bloat the feature set of the
app by providing multiple solutions to a single problem... all just to
satisfy one user who is too stubborn to even think that his method might
just not be the best.

When you start adding donations, things become EXTREMELY politically
charged.  Users then EXPECT that a feature they donated to be
implemented.  Depending on how the donations work, which country the
donater is in, and so, you could even end up with developers being
legally obliged to implement an suboptimal or poorly thought out
feature.

In general, I like the idea of being able to donate money to help out
with bounties (i.e., requests for features and bug fixes.)  However,
users SHOULD NOT EVER be able to decide which bounties are available.
The developers, product managers, QA directors, and other similar people
should work together to decide on the available bounties.  They can weed
out bad requests, work to propose alternative solutions, and so on.
Once the general approach is determined, the bounty can be placed and
users could donate, such that the bounty is not ever a request for
something the project doesn't want to do, but instead serves solely as a
way to garner developer interest in a particular feature or bug that
otherwise wouldn't get implemented due to lack of time or interest.
That feature will still be something thought out and agreed upon by the
project maintainers, though.

Right now, a new feature often gets implemented only if the requester
posts a patch.  Even *with* a patch, though, bad features can and are
rejected.  By going with a bounty-based system, you ensure that
maintainers can still reject those bad features (by never placing a
bounty) but outside developers can still submit patches, and that both
the core maintainers and outside developers can be motivated to work on
a difficult or "unsexy" feature.
-- 
Sean Middleditch <elanthis awesomeplay com>