wiki:SprintPlanning20111025

BIND 10 Sprint Planning meeting, for the sprint ending 2011-11-08

Held on 2011-10-25 14:30 UTC

Attendees

Aharen
Kevin
Fujiwara
Stephen
Jelte
Michal
Jeremy
Jinmei

Summary of past sprint

  • Number of closed tickets: 15
  • Total estimation for closed tickets: 45 :(
  • Tickets still new in past sprint: 8
  • Total estimated time for those: 31

Past Sprint review

jelte: we didn't get a lot done in terms of estimation points. Almost 30 points are in review or ready to merge.

Stephen: suggest that we also report how much time we spend on BIND 10 as soem of us spend time on other things (e.g. BIND 10 DHCP).

Jelte: spend a lot of time on system testing and has not finished a ticket.

Vorner: also spent time on a large task without finishing it.

Jinmei: this seems to be a common problem if we have large tasks, and i don't see a good solution. We could be careful about task sizes in the future but we cannot completely eliminate large tasks.

Stephen: that was why i used a moving average

Jinmei: it also seems a number of people have been busy with other things

There seem to be a number of tickets that are either ready for merge or only waiting for a final ack, but even then we did not really reach our normal productivity (as far as we can tell with the estimation process). Any improvement points?

Jinmei: maybe we have too many different things to work on, and we didn't concentrate on one specific feature etc.

Stephen: if we have just one feature, tasks will get very tightly coupled, and we may be blocking each other. The other thing is that if you have multiple people working on the same feature, you need to specify it really well, so the different parts integrate well.

Jinmei: i did not necessarily say it was a bad thing, it is just an observation did we reach our goals?

  • Design for data source differences

Basically done, based on discussions on list.

  • (design for) handling of DDNS commands

Nothing done on this

  • Depending on whether the current state is advanced enough, we may want to finish boss configuration.

Not finished, three pages of comments, michal is about to answer the last ones after this call. After that there will probably be a few more iterations. Also a number of followup tickets in proposed queue already.

New sprint goals

proposal:

  • Start with ixfr-out implementation
  • Start with DDNS design
  • continue/finish(?) ds refactor

Is this too ambitious? Not enough? Just right?

Jinmei: we can complete the difference framework, but not complete ixfrout.

Changed first to:

  • Finish the differences framework implementation

Stephen: how do we actually generate the differences?

Jinmei: we get diffs from other server for now

Stephen: so we can store and send them, just not create them ourselves?

Jinmei: yes, if we want and have time, we can work on generate-from-differences

Stephen: How does this relate to the discussion we had in amsterdam, about multiple versions of zones and how they are mapped in the data source?

Michal: that is different from this work

jinmei: We can include task breakdown for DDNS feature, personally I don't think we need to worry about design for this too much.

Jinmei: should we focus on getting ddns in next release?

Jelte: depends on the result of this work imo

Jinmei: I think we can get ixfr-out in, but probably not full ddns support

What bout ds refactor?

Jinmei: i think we can complete NSEC part, but not NSEC3. Original plan was to break down NSEC3 into smaller tasks, but that has not been finished. A more reasonable goal is to put the breakdown task into the coming sprint. Hopefully we can complete the refactoring in the next sprint then.

Jelte: this has taken way too long already

Jinmei: from the user point of view it doesn't really much that it is not completely done, since we can still work on the new stuff while the auth uses the old interface.

Michal: I'd like to see a real database like mysql instead of sqlite3, which would be a user-visible feature.

Jinmei: not so sure about performance of that, but supporting some more database backends will definitely be attractive for users.

Stephen: I think we need to focus on our year-end goals here.

Ticket carry-over

What to do with #1092?

Jinmei: my suggestion is to exclude it for the next sprint, and move it from the review queue unless it is urgent.

Jelte: I'll not move it to next sprint but will take it up with Likun

#1166

Jeremy: I made some changes and stephen OK'd them, but then he made more changes and those need to be reviewed.

Jinmei: can we merge the completed part and create a separate ticket for the rest?

Jeremy: The goal is to be able to run doxygen without any errors, we need something to force us to keep doxygen generation clean, but now more and more errors appears

Jinmei: then we should probably not try to complete all errors in one ticket, but make progress step by step, and then include at least one the subtasks for each sprint (similar to defect tickets)

Do we need followup on #1279? (e.g. differentiate between errors that should cause fallback and those that should not)

The proposal is to merge it behaving as it does now, and have a discussion about whether to extend it on -dev. Certainly not adding tickets for such a thing at this moment.

  • Total of 37 points by carry-over

Proposed ticket queue

Taking from the proposed queue:

  • 1329-1333 (for differences goal)
  • 1347 (for ddns goal)
  • 1343-1345 (as defects)
  • 1028 (may or may not be a lot of work, unknown at this point)

Any other tickets

No

Enough defects

2 carried over, 4 from proposed.

Statistics tickets?

Aharen: two tickets assigned to me, i will continue 509 and 510, kambe also has two tickets, 1275 has been done, and 917 is open, he will work on those in this sprint.

Jinmei: should we do estimation and estimation review on the statistics tickets? It's a half-blackbox, partly because it is not explicitely included in the planning and review process. I propose that even if we count them separately I propose the jprs submit their plans to jelte or the team or dev list, and we see its progress.

Aharen: the tickets are quite big, and not broken down, so estimation of amount of work may be difficult. We can try to do estimates for future tickets

Jinmei: Oh yes I was not saying anything should be done for this past sprint, but perhaps we can do something in the future

Jelte: we should certainly try to keep statistics tasks small as well, for better overview.

Fujiwara: We will have a discussion about this within JPRS next week, and we will report back

Last modified 6 years ago Last modified on Oct 25, 2011, 4:31:12 PM