Conversion of Priority field values (added to task description on 2014-07-23):
Bugzilla | Phabricator |
Unprioritized | Needs Triage |
Immediate | Unbreak Now! |
Highest | Unbreak Now! |
High | High |
Normal | Medium |
Low | Low |
Lowest | Needs Volunteer |
Everything below only for historical reasons.
Andre's initial rough proposal to iterate on, plus some thoughts:
Bugzilla Priority | Phabricator Priority | Comments |
Immediate | Unbreak Now! | |
Unprioritized | Needs Triage | |
Highest | Drop. in a Phab world, this should be covered by making the task a member of a sprint project, I think? | |
High | High | |
Normal | Medium | "Normal" is subjective, cf. https://bugzilla.wikimedia.org/show_bug.cgi?id=46020 |
Low | Low | Or does "Low" sound too derogative? "Far future"? |
Lowest | Low | Or does "Low" sound too derogative? "Far future"? |
Wishlist | I am not very happy with this and it depends highly what "Wishlist" means: Functional feature request vs. I'd wish this bug got fixed? Any ideas? Enhancement requests are independent from Priority (Phab default implementation) and Severity (Bugzilla default implementation), IMHO, plus feels very similar to low priority by default anyway. What is the actual need to differentiate between feature requests and bugs? Pure statistics and influence on burndown statistics? | |
Global meaning of values: There are some Bugzillas out there using generic priorities like P-1 to P-5 (leaving exact interpretation to teams), but I'd prefer to aim for a shared and global understanding because "outsiders" with their expectations wouldn't need to look up unwritten interpretations of each team.
Why potential drop highest priority: Bugzilla gives you a group of tickets with priority XY, but inside of that group it is currently impossible to make a ranking - which one is first? (cf. page 297 of http://grouplab.cpsc.ucalgary.ca/grouplab/uploads/Publications/Publications/2009-IssueTracking.Report2009-933-12.pdf ). Phabricator will fix this by providing workboards, so I have the naive hope that we'll need less priority values.
Why trying to reduced the number: Though it is not possible to directly transfer findings in other projects / audiences, in Eclipse Bugzilla, "only the second top developer is using the five levels of priority. The rest are using only two or three levels.", plus "While developers are using only three priorities, not all developers are using them in a consistent manner. Some never use priority 1 and 2, and some never use 4 and 5." quoted from http://herraiz.org/papers/english/msr84cp-herraiz.pdf pages 147-148.
(Nota Bene: Those researchers only analyzed ''closed bugs'' in Eclipse Bugzilla but did not clarify if that means FIXED or also other resolutions, plus it is based on the highly questionable assumption that the fields are correctly used, but fixing bug metadata after resolving an issue normally does not take place because nobody cares - see http://www.st.cs.uni-saarland.de/softevo/bugclassify/paper/icse2013-bugclassify.pdf page 396)