Hi Kathey,
Thanks for all of the time which you have spent cleaning up the bug
reports recently. Let me give my $0.02:
1) I'm not sure what kind of prep can be expected before this meeting.
In the time between now and that meeting, the Sun engineers are going to
be busy at Java and Community One.
2) We have a lot of fields on our JIRA reports. The following three
fields confuse me because I don't see any consistent pattern to what
they mean. Their contents seem to be subjective. Because they are
subjective, they don't give me much guidance in selecting bugs which
should be fixed:
Priority
Urgency
High Value Fix
3) The following fields could have rigorous meaning and would be useful
to me in choosing bugs. But they have limited value right now because
people don't always remember to fill them in. It might be worthwhile to
spend a day or two agreeing on what these fields mean:
Crash
Data corruption
Deviation from standard
Performance
Security
Wrong query result
Regression
4) In addition, I am influenced by the number of Votes which a bug has
received.
5) Before calling a meeting, I would like us to triage the 355 or so
open bugs. That is, I would like us to make sure that we have filled in
the fields listed above in item (3). This is a job which we could divide
up among all of the willing engineers on this list. We could coordinate
via email or a wiki page. I would certainly be willing to triage a block
of 50 bugs--I would probably be dismayed by the prospect of having to
triage all 355 bugs by myself.
If we then ask the users to vote on the triaged bugs, I think that I
will have all the information I need in order to focus my attention
during the 10.5.2 bug-fixing phase. I'm not sure what I can contribute
to this meeting if the bugs have not been triaged first. I'm also not
sure what additional guidance this meeting can contribute once the bugs
have been triaged and the users have voted.
Just my thoughts.
Thanks,
-Rick
Kathey Marsden wrote:
I would like to plan a bug review meeting on #derby IRC on Monday,
June 8 at 15:00 UTC (that's 8:00am PDT), (5:00pm CEST) if I figure
this correctly.
Since we have a lot of open issues, I would like to focus this meeting
on the High Value Fix Candidate list:
https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&&pid=10594&customfield_12310200=High+Value+Fix&resolution=-1&sorter/field=issuekey&sorter/order=DESC
The goal of the meeting will be to get this list down to a list of
40 issues that the community agrees will offer highest return on time
investment. Currently the list is at 53. Some will come off and I am
sure others will be added as part of this process.
The way the game works is *before* the meeting (by Friday June 5)
everyone goes through whatever lists they like and marks the open
issues they think belong on the list by checking the checkbox in
Jira. You may find the reports on this page useful:
http://wiki.apache.org/db-derby/DerbyJiraReports. If you added an
issue to the list, you can take it off, but you can't take off an
issue someone else marked high value, without getting their ok by
posting your request to have it removed to the Jira issue.
Come to the meeting ready to advocate for your issues. At the meeting,
we go over the high value list ,and trim it down to 40. Bumped issues
will get a comment in Jira with an explanation and of course can go
back on the list once we clear the backlog a bit.
Does anyone have any objections to the format of the meeting?
Will someone volunteer to be secretary and summarize to the list and
make the agreed upon Jira modifications?
Also please look at the Wiki page definition of High Value
http://wiki.apache.org/db-derby/HighValueFixCandidates
Questions:
Is this definition ok or could it be made clearer?
Should the list include improvements and testing build tool
issues or just bugs?
Should we include documentation? I think DERBY-4165, DERBY-4034,
and DERBY-1209 would make nice additions, but of course that eats into
our 40.
Thanks
Kathey
