Re: Minutes of the Board meeting of August 29th, 2014



Their terms of service [1] are also highly questionable. They reserve
the right to take all the money in our account, or cancel our bounties
without refunding our money, at any time and without any explanation,
without cause or notice (section 6.5). "Without cause" seems like a
pretty good place for us to draw a line....

Daniel Foré should probably be the one who we should alert.

OK, I've CCed Daniel to make sure he's aware about the issues with
Bountysource's ToS. It's probably not a huge risk, but it would not be
fun to be wrong about that....

We don't
use it ourselves.  However, I am intrigued if people believe that we
should put out bounties?  What I get a lot of times and you've seen
this yourselves is that a lot of bugs are complex and takes a lot of
intimate knowledge of the GTK+ or GNOME stack.  Would it be effective?

sri

I think it could be effective, but it's something that would really
require strategic discussion to decide which bugs to place bounties on,
and how large the bounty. Presumably we would get better at this with
experience. I suspect that the most effective approach would be to
identify high-profile bugs that are not attracting sufficient developer
attention (e.g. the bad search experience in nautilus), and try to guess
the appropriate amount of money that will incentivize work on the bug.

Adam's $400 bounty for what appeared to be a simple crash actually
required digging into the webkitgtk+ sandboxing code, which basically
limited the bug to those who already have at least some webkit
experience. It did work in the end when it attracted my attention, but
that took a while.

My $150 bounty on the gnome-shell desktop background bug hasn't been
successful yet; it probably wasn't large enough to motivate anyone
unfamiliar with gnome-shell to look into the problem. I figured a
smaller amount of money would be appropriate since it's probably a
fairly simple issue and I pointed out a commit that I thought was wrong.
Maybe someone will work on it in the end, but I have a feeling it will
end up being solved by an experienced gnome-shell developer, and I guess
the bounty will probably not be significant motivation for doing so.

My $250 bounty on the gnome-shell calendar bug seems to have been just
right, though: that bug wasn't getting any attention and wouldn't have
been fixed without the bounty, and now it's very likely to be resolved
soon.

Attachment: signature.asc
Description: This is a digitally signed message part



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]