Re: GNOME, Bounties and paid development [Was: Re: OPW; Where does the 500$ for each GSoC goes?]

[Germán Poo-Caamaño <gpoo gnome org>]

[Resending in full, as in my previous I accidentally pressed the Send button by
mistake]

On Wed, 2014-09-17 at 15:43 -0500, meg ford wrote:
> On Wed, Sep 17, 2014 at 2:08 PM, Tristan Van Berkom <
> tristan upstairslabs com> wrote:
> > I think perhaps, if we organized bounties which clearly and definitely
> > improve software that industry is going to use, and not only for the
> > singular purpose of the GNOME Desktop Environment, then perhaps we would
> > be able to get some real backers in the industry to come together with
> > us and put together a bounty that is worth bidding for.
>
> I think there are two ways to approach this: (1) the way you suggest above;
> and (2) by having smaller bounties which do not require a bidding process
> and can be picked up by contributors who would generally donate time, but
> could use some extra money in order to afford to contribute their time.

Either way, there is a high administrative overhead, as it was proven in
the past when Novell (in communication with GNOME Foundation), and then
GNOME Foundation with the help of Google tried bounties. All of them for
GNOME.

Before going to some details, the short story is: the outcome was what
we know now as Google Summer of Code.

I don't pretend anyone stop pursuing bounties on GNOME, but please,
don't start from scratch, dig the board meetings, the mailing list
archives, and any other source of information.  This has been discussed
lengthly in the past.

Some of the administrative things you have to consider:
      * Determine the issues you want to fix, assess them and put them
        value.
      * Set the rules
      * Getting the maintainers involved to: (1) check if the fix is
        worth, (2) review the patches, (3) pushing them in master
      * People to track the patches, update status of bounties (to avoid
        double work for potential contributors)
      * All the dance to exclude people from certain places where GNOME
        Foundation cannot send money to, get bank information, wire
        money, track everything was good, and makes that everything goes
        well that IRS won't complain.
      * ...

Also, what is the goal you want to pursue? Who would you expect to apply
for?

Don't even mention that for maintainers will have to spend some time for
answering questions, to answer how the things work they way the do,
things they don't want to see in their projects, and so on.

I quote an email of Nat (somehow forwarded to wikimedia foundation):

        [...] One quick point on numbers: only 11 bounties have been
        paid, but we've had patch submissions on >50% of the total
        bounties; release engineering timelines have made it hard for
        bounty submitters to get some of their patches accepted by
        module maintainers, and therefore paid, so that contributes to
        the small number of paid bounties you see.
        
        One thing that's surprising is that pretty much all of our
        bounty
        submissions came from first-world economies.  Despite efforts to
        promote the bounties heavily in e.g. India.
        
        I think there's a need for a bounty administration
        infrastructure; some piece of software that can run these
        programs automatically, instead of the mostly hand-generated web
        pages I wrote.

        http://comments.gmane.org/gmane.org.wikimedia.foundation/797


Current board members can dig in the board-list archives from
2003-2006-ish, to get more details.  I bet there are also here in this
list, and in the board meeting minutes, for example, from 2004-08:
        
        Nat gave a summary of the current situation with the Bounties.
        We've reached the 1st round and so far give out $7460 for 10
        bounties. It's currently stalled because of an Evolution code
        freeze - will launch again soon with a new deadline. The
        Bounties are proving to be effective in attracting new Evolution
        developers. Nat asked if it would be useful to have a general
        mechanism where anybody could put money against a bug item? Owen
        asked if it was more interesting maintaining a TODO list. Luis
        worried that making the TODO list the Bountie list was
        dangerous, because people might end up doing only the things
        people pay for. Have we already started down this slop already
        with company involvement?

        https://mail.gnome.org/archives/foundation-list/2004-August/msg00173.html

I think Luis Villa's concern still is valid.

The first try was "Desktop Integration Bounty Hunt", funded by Novell
but "lead" by GNOME Foundation:
http://web.archive.org/web/20070210190246/http://www.gnome.org/bounties/

The next time, it was funded by Google, applying some experience from
the previous bounty hunt:
http://web.archive.org/web/20070208175703/http://www.gnome.org/bounties/Google.html

This was before GSoC exists, but it starts to resemble it. 

There was even a wiki page to discuss lesson learn, in this case, mostly
for GIMP bounties:
https://wiki.gnome.org/Attic/BountiesDiscussion

-- 
Germán Poo-Caamaño
http://calcifer.org/