Blocker process: tracker bug / whiteboard naming proposal

Adam Williamson awilliam at
Sun Jan 20 18:49:16 UTC 2013

Hi folks! We're running a FUDCon NA workshop session today to try and
actually knock out some QA 'low-hanging fruit', and the one we tried to
deal with this morning is the blocker tracker bug names. We had some
grand ideas for making this even better in the glorious future, which
involve more good stuff in the blocker webapp, but we intentionally
restricted ourselves to the scope of changes we could propose and
implement quickly, for concrete improvements in the Fedora 19 cycle.

This is an issue that's come up repeatedly in the past and was also came
up during the talk Tim gave on the blocker process on Friday: the
feedback from everyone at the meeting (QA people and non-QA people) was
that the current names were really bad and confusing. So we're
definitely addressing a need here.

What we (Tim, John Dulaney and myself are here) came up with is a simple
renaming of the aliases for the blocker / NTH tracker bugs and the
whiteboard fields used to track proposed/accepted state, proposed by
Tim, that addresses a lot of the problems that just about everyone has
noticed with the current names:

* The word 'accepted' is used in two different contexts in the same
process, which confuses everyone
* The format of the aliases is consistent for Alpha and Beta but not for
Final, so you can't just remember them simply
* 'Nice to have' is a confusing term - it makes sense from the
perspective of a person sitting in a blocker meeting, but it's confusing
in other contexts, and not clear what it really means

We also noted as a problem that 'it's not obvious to someone who doesn't
know the process that a proposed blocker is only proposed, not accepted'
but came to the conclusion that this is not fixable within the narrow
scope we set for this proposal - to fix that we'd have to have more
tracker bugs, or the system Tim's working on for tracking blocker state
outside of Bugzilla.

The proposal is this:



This could be 'camel cased' (mixed upper and lower case) as the old
names were, or not - camel casing makes the separate words clearer, but
makes people think the case is important, which it really isn't
(Bugzilla is case-insensitive in actually handling aliases; marking a
bug as blocking 'f18alpha' works fine, for instance, you don't have to
type 'F18Alpha'). It's not hugely important either way.

There are really just two changes here: make final consistent with alpha
and beta, so that the phase name is in the aliases, and use
'freezebreak' / 'fb' instead of 'accepted' / 'NTH'. The word
'freezebreak' is clear and unambiguous and, as a bonus, functionally
descriptive: the whole purpose of this process is to track and review
freeze breaks. That change also means we are not using the word
'accepted' in the alias names any more. Putting the phase name into the
aliases for Final is a break with tradition (back in the mists of time
before Alpha and Beta blockers were really tracked and before the
NTH/freezebreak process existed, we just had a single bug named
'FXBlocker' for each release), but it makes the names consistent and
hence memorable - right now people always get tripped up trying to do a
simple 'if it's X for Beta it must be Y for Final' operation.

The nice thing about aliases is that a bug can have multiple aliases, so
our proposal is that we edit the bugs for F19 to have both the 'old' and
'proposed new' aliases, and started using 'acceptedfb' instead of
'acceptednth'. We try and get used to using the new names in blocker
meetings, casual discussion etc. Then, if all goes well, we just have
the new names for F20. We would also have to edit the relevant SOP pages
to reflect the new nomenclature, obviously, but that's pretty much a
search/replace substitution and doesn't need to be drafted and proposed
separately - if this proposal is accepted, I'll just go ahead and make
the appropriate wiki edits.
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Twitter: AdamW_Fedora | adamwfedora

More information about the test mailing list