[Linux-aus] LCA2016 post event report

[Jonathan Woithe]

On Sun, May 29, 2016 at 05:36:56PM +1000, Kathy Reid wrote:
> For the Miniconfs at LCA2016, Schedule changes were being updated in four
> primary locations;
> 
> - The Miniconf site by the Miniconf Organiser
> - The Wiki page by the Miniconf Organiser
> - The ZooKeepr schedule by me
> - The mobile app by whoever wasn't tied up with a higher priority task
> :
> So, in the event that a Miniconf Schedule changed, up to four people were
> making an edit to four sources.
> :
> Definitely recommend synchronising Miniconf Schedule timeslots, times, and
> reducing changes.

Coming up with a way to streamline the distribution of miniconf schedules
would definitely be useful.

It would be good if we could come up with a consistent approach which
applied across all miniconfs.  This would not only potentially reduce the
amount of maintenance effort required but would also be less confusing for
delegates.  Since AV utilise zookeepr schedules to coordinate their thing,
the adopted approach probably has to involve that.  If this were the case,
it would be advantageous if the miniconf organisers themselves could manage
their miniconf schedule within that system, thereby freeing the LCA
organisers from this task.  If there was a single definitive source for
miniconf schedules, anything else which needed schedule information could
pull it from that.  Having said that, I understand that manipulating
schedules within zookeepr is not straight forward, so some work might be
needed in this area for such an idea to work.[1]

The nature of miniconfs is that schedule changes are a fact of life and
often these occur quite late through no fault of the miniconf organiser. 
Making schedule changes less onerous is therefore a good goal, even if it's
something which will necessarily take some time to implement.

Regarding coordinated schedules (a separate issue), in principle I have no
problem with this although I acknowledge that for some miniconfs it won't
work very well.  As a result, we need to be flexible when necessary.  In any
case, if a coordinated timetable is desired I think many miniconf organisers
would like a definitive indication of this as early as possible, along with
the preferred timetable.  This means they can plan the talk slots and their
miniconf generally around the published schedule from the very start.  Once
one has a few speakers locked in it is often difficult to rearrange things
to match up with a subsequently published schedule.

When I published my schedule for LCA2016 there was only one other miniconf
schedule published for that day.  In the absence of any other guidance, I
aligned with their schedule because it was trivial for me to do so.  When
other schedules appeared they did not align, but this was unsurprising
because they had obviously worked to a different schedule and changing
things at that late stage was not going to be feasible.

>From my perspective as a miniconf organiser, coordinated miniconfs are a
good thing to aim for.  However, to make this work the following will be
required:

 1) Notification from LCA organisers that coordinated schedules are a thing
    as early as possible.

 2) Early distribution of the miniconf time slot schedule to organisers so
    it's available to them before CfPs are issued.

 3) Allowances for those miniconfs whose content does not lend itself to
    a traditional schedule and might be difficult to fit around the
    published schedule.

None of these steps would add to the workload of LCA organisers because they
are largely policy decisions.

As to what a proposed schedule might look like, I agree with others who have
cautioned about loosing too much time to breaks.  Unless miniconf rooms are
widely separated from each other, a 5 minute break is probably sufficient. 
If someone is concerned about not making it they could always depart during
question time before the break.  As to how often a 5 minute break should be
scheduled, I'm not overly fussed.  Personally I think a coordinated break
after 50-ish minutes is probably workable.  Doing it after only 20 or so
minutes might be a bit much as it will be limiting for those miniconfs whose
speakers tend to only need 10 minutes to present their talk.

Based on the session times from LCA2016, something like this might work:

Session 1: 10:40 - 12:20

  10:40 - 11:30  = 50 minute block

  11:30 - 11:35  = Coordinated Break

  11:35 - 12:20  = 45 minute block

Lunch: 12:20 - 13:20

Session 2: 13:20 - 15:00

  13:20 - 14:10  = 50 minute block

  14:10 - 14:15  = Coordinated Break

  14:15 - 15:00  = 45 minute block

Afternoon tea: 15:00 - 15:40

Session 3: 15:40 - 17:20

  15:40 - 16:30  = 50 minute block

  16:30 - 16:35  = Coordinated Break

  16:35 - 17:20  = 45 minute block

In addition and if deemed necessary, there could be additional *optional*
coordinated breaks within the blocks.  That is, if for a given miniconf
there is a natural break between speakers in these blocks, then those breaks
could be scheduled in a coordinated way (say, 20 minutes in).  However,
these breaks will create problems for some miniconfs so they would be seen
as being optional.  Those who want to break the blocks should do it then,
but otherwise don't bother.  This provides the flexibility to allow for
different types of miniconfs (for some it makes sense to have one or two
speakers per block, while for others with shorter presentations it's
appropriate to have 3 or 4 per block).

The above schedule attempts to balance the concerns raised: having some form
of coordinated break while not having these consume a considerable
proportion of the miniconf's total time.  The final outcome may look nothing
like the above but regardless of the final approach adopted, keeping this
balance in mind is important.

Regards
  jonathan

[1] I have never worked with zookeepr, so the comments about managing time
    schedules within zookeepr are based solely on public statements made by
    others over the years.