On Tue, Apr 2, 2013 at 6:53 PM, Alex Huang <alex.hu...@citrix.com> wrote: > So let me start off with I agree in principle with what Noah is talking about > here. Cookie licking is an anti-pattern that we should reject as a > community. However, I disagree the solution or even what is perceived as > cookie licking. > > We established a while ago as a community that we follow a Jira workflow in > handling bugs. In that process, a bug is Open until it goes to "In > Progress". We can simply tweak our expectations on this workflow to achieve > exactly the desired effects. > > - Assign bugs does not mean you own it, even if you assign it to yourself > doesn't mean you own it. > - No one owns a bug until the person assigned changed it to In Progress (not > licked til this stage) > - Before a bug goes into In Progress, anyone can grab it. > > By doing this we allow people who can help in prioritizing the bugs to assign > bugs without going through another layer of negotiation. Assigning bugs > merely means asking the question "can you work on it". This would be much > more efficient way of doing things. > > I also think that we can setup public filters that people can use to find > bugs they can work on. In the filter, it doesn't look at who the bug is > assigned to. Just whether the bug is open or not. > > --Alex
I understand the reasoning - but for a newcomer looking to get involved, I think 'assigning' a bug - whether by default, or otherwise can be construed as excluding newcomers and no room for them to get involved, so I think it warrants caution at a minimum. Our 'if not 'in progress' anyone can grab it' does not seem to be the norm for open source projects. I also wonder if it makes a difference. Perhaps it does. In other projects I've been involved in, when I care about releases or just the project, I tend to watch the categories of bugs that I can fix, particularly the unassigned items. Does having a backlog of work assigned to me (as opposed to checking a component in Jira for outstanding work) improve things somehow? I don't know, perhaps it does and there is something I am missing. I do think triaging is important - there are simply too many bugs to not have some ongoing triage, make sure the proper component, severity, etc is set, perhaps even doing some initial prompting for enough information. --David