On Wed, Apr 13, 2011 at 12:44 PM, Krinkle krinklemail@gmail.com wrote:
I agree. Defaulting new bugs to a low priority doesn't seem very friendly to new users. They don't know (and shouldn't have to know) what the bugmeister's organization is.
I thought about replying with a similar response, but then I realized that if Mark stays on top of incoming bugs (which is his job), he'll be able to bump bugs up to normal as they come in, which should actually make those people good, and provide some automatic positive reinforcement for filing important bugs. I suspect most people won't even notice that the bug defaults to "low", and they'll be able to fix it if they disagree. I suppose the best thing would be to have a blank priority (or "triage" as OQ suggests), so that we could query for the bugs that haven't been explicitly prioritized, but I don't know if that's possible.
This isn't so narrowly focused on the bugmeister as it is starting to pay attention to the priority field as a community. All priorities should have a meaning for us. We can debate about what is "correct", but here's what my thinking is:
1. "Highest" - this means a bug that someone is either working on *right now*, or at least should be working on right now. No developer should have more than one "highest" priority bug, and anything marked "highest" priority shouldn't stay unassigned for more than a week. 2. "High" - this means that a bug is in the relatively-small pool of bugs that will likely get bumped to "highest" priority. Bugs at this priority should, at a minimum, be blockers for the next release of the software. 3. "Normal" - these are bugs that really should get fixed for the next release of the software. They aren't necessarily blockers, and we may grit our teeth and have a release or two with this bug, but there's no way we should go two releases with a "normal" priority bug. 4. "Low" - these are bugs that we won't actively manage. They may be important to some people, and someone may get around to fixing the problem, but no developer is committed to working on a fix 5. "Lowest" - these are bugs that pretty much everyone agrees aren't terribly important (including the reporter).
I'm not wedded to these exact definitions, but I'm pretty wedded to the idea that we should agree on something. I don't think we should consider it "normal" to have bugs go for years without activity.
From a practical perspective, many more bugs than are being marked as
"normal" are actually "normal" by this standard. Most of the bugs in our database are probably "low" priority in the sense that we just can't get around to fixing all of them. I can see the argument for making the most common case the default (though I share others uneasiness with doing so).
Rob
On Wed, Apr 13, 2011 at 1:15 PM, Rob Lanphier robla@robla.net wrote:
From a practical perspective, many more bugs than are being marked as "normal" are actually "normal" by this standard. Most of the bugs in our database are probably "low" priority in the sense that we just can't get around to fixing all of them. I can see the argument for making the most common case the default (though I share others uneasiness with doing so).
It may be worth having an "unassigned" initial priority, which we can treat as equivalent to "low".
-- brion
Brion Vibber brion@pobox.com writes:
On Wed, Apr 13, 2011 at 1:15 PM, Rob Lanphier robla@robla.net wrote:
From a practical perspective, many more bugs than are being marked as "normal" are actually "normal" by this standard. Most of the bugs in our database are probably "low" priority in the sense that we just can't get around to fixing all of them. I can see the argument for making the most common case the default (though I share others uneasiness with doing so).
It may be worth having an "unassigned" initial priority, which we can treat as equivalent to "low".
We can do this, I think, by changing the bug's state from “NEW” to “ASSIGNED”.
Mark.
From a practical perspective, many more bugs than are being marked as "normal" are actually "normal" by this standard. Most of the bugs in our database are probably "low" priority in the sense that we just can't get around to fixing all of them. I can see the argument for making the most common case the default (though I share others uneasiness with doing so).
A good thing is that if the default is normal and the bugmeister then lowers it, that gives a negative feedback to the reporter, whereas if it's as the default, it doesn't.
Most bugs that are bugs should be normal, but perhaps we don't have the resources to fix each of them.
wikitech-l@lists.wikimedia.org