launchpad-dev team mailing list archive

[Robert Collins <robertc@xxxxxxxxxxxxxxxxx>]

tl;dr:
 * expect about 600 more bug mails
 * tell me about any themes or sorts of bugs that you want to see kept
in the high bug list

So, in the last overhaul of our bug triage policy, we agreed that we'd
regularly do a sweep of the high bugs and keep them under control -
there should be only 6 months worth of bugs there (which is 500 for
maintenance squads @ 25/week + 100 or so feature bugs which are quite
a bit heavier).

With our mediums overhauled - and that felt really good, there was -so
much- crud in there that was already fixed or now irrelevant - I'm
going to tackle the overdue review of High bugs.

This means that there will be about 600 bugs, or 50% of the high bugs,
that we need to say 'this is less important than the rest of the
highs', which is a lot of bugmail :(.

Now, we might argue that this isn't worth doing because with 260
criticals we don't have highs being worked on anyway, so 'meh' - but
thats too simple a view IMO: the feature squads are still working on
high bugs, and the overlarge set of highs impacts them to a degree.
Secondly, the whole point of doing an assessment of highs is that
unlike criticals we have no objective rules to instantly categorise
everything that should be high: sure we have those bugs for features
which are in the queue to be worked on, but beyond that high merely
represents our collective opinion about things we'd like to do very
soon (be it a pet bug, a feature request for Ubuntu, or UI glitches
that are just so annoying!). This is a crucial difference between the
critical and high buckets, and regardless of what we do with criticals
that difference will remain: we'll have a large number of bugs which
we want to work on for primarily subjective reasons, and we need to
keep that list under control (or alternatively not have bug importance
at all - a valid choice but not our current policy :)).

If the record we had with mediums holds true, assessing all 1200 high
bugs will close a significant number (possibly a hundred or more) but
beyond that its going to be hard slog deciding which of them is more
urgent; so I expect I will make some mistakes and unqueue something
that truely needs to be in the 'next 6 month queue' : if I do that,
please just pop it back to high - I won't argue... but I will ask that
if there is more than one or two such bugs you take the time to find
and unqueue a couple of the other highs.

I've spoken to Matt and gotten a list of feature work bugs to look out
for; do you have any pet things that I should be holding dear while I
do this broad assessment? E.g. ie support, or js support, or
performance...

Cheers,
Rob