Re: [Debconf-team] Accommodation numbers/upgrades/communal accommodation
Hi
Steve Langasek <vorlon@debian.org> writes:
> Hi Gaudenz,
>
> On Mon, May 20, 2013 at 10:29:52PM +0200, Gaudenz Steinlin wrote:
>> >> We are now (by tomorrow) at the end of the period for applying for
>> >> sponsorship, so the next set of decisions about room utilisation etc.
>> >> should be made soon.
>
>> >> How many people have registered for what room categories currently, and
>> >> how does that compare to the available rooms? How many sponsored people
>> >> said they would/wouldn't take "communal" accommodation?
>
>> > The room categories are visible in this munin graph [1]. Additional data
>> > about the "Debian role" is available in the admin interface.
>
>> > Current status:
>> > All / accompany / interested / None / not selected
>> > Own accomodation: 26 / 0 / 10 / 3 / 0
>> > Sponsored accomodation: 121 / 0 / 5 / 0 / 0
>> > self-paid communal: 16 / 0 / 3 / 1 / 1
>> > self-paid camping: 13 / 1 / 0 / 0 / 0
>> > (+ 20 requesting sponsored accomodation and prefering camping)
>> > self-paid 8 bed: 27 / 1 / 3 / 1 / 0
>> > self-paid 2 bed: 27 / 0 / 2 / 0 / 1
>
>> > Not taking into account any additional requirements this means we have
>> > enough rooms for all these requests.
>
>> >> We should clarify soon for people already registered what kind of room
>> >> they can expect to receive, and, e.g. give them the chance to pay for a
>> >> better room if they aren't happy with that one.
>
>> > I think we should as soon as possible confirm the rooms for those that
>> > choose a self-paid room. For the upgrades we have first to find a more
>> > or less consensual way of doing this. As a discussion basis for this we
>> > need to know the answers for the survey questions. AFAICS these are
>> > currently not accessible in a munin graph or in the admin interface. As
>> > it looks now, we still have beds available in each category.
>
>> I just got read access to the penta DB and tried to retrive the data
>> about the communal accomodation survey [1]:
>
>> 95 would accept communal accomodation
>> 17 would *not* accept
>
> Thanks for running this query. It seems that the models advanced for
> sponsored attendees by the local team are reasonably accurate, then, and
> there is no need for major adjustments to the undersubscription of the
> rooms.
>
> Just a couple of things:
>
> - The early query showed a total of 121 people requesting sponsored
> accomodation, and your query only includes 95+17=112 of these people. We
> should understand where the other 9 people are. Have they requested
> sponsored accomodation, but failed to mark "I want to attend this
> conference"? Or are they missing for another reason? If people failed to
> mark "I want to attend", we should probably consider that a schema bug
> and include such people in the count of possible attendees.
This is because of two things: First 7 people did not answer the
question. Second the number of 121 above includes 2 persons who did not
check the "I want to attend" box.
Here is more detailed data also including the camping question:
communal | camping | count |
-----------+---------+-------+
f | f | 16 |
f | t | 1 |
t | f | 79 |
t | t | 16 |
| t | 3 |
| f | 4 |
> - I'm very confused by what I find at
> <https://wiki.debconf.org/wiki/DebConf13/Pricing>. I understood that
> there was very strong opposition to filling the largest rooms to their
> nominal capacity; but this chart shows all rooms with 8 or more beds
> being filled to capacity. And looking at
> <https://wiki.debconf.org/wiki/DebConf13/Pricing/Compromise>, I see it
> redirects back to <https://wiki.debconf.org/wiki/DebConf13/Pricing>...
> and that the last version of /Compromise matches. So I'm concerned that
> this page doesn't actually represent a shared understanding of how the
> rooms will be allocated. Could someone please either confirm that we
> *are* planning to fill these large rooms, or else fix /Pricing to reflect
> the actual plan?
As far as I recall we decided to not fill these rooms. The meeting
minutes of the IRC meeting on 8th April seem to somwhat confirm this.
But unfortunately nobody did a "agreed" meetbot command about this, so
it's not super clear. [1]
But the Pricing page did not get updated after that meeting. So IMO the
wiki page is wrong about this.
>
> That looks like:
>
> - 164 attendees (sponsored+non-sponsored) in rooms >=3 people
> - 63 sponsored attendees in the reserved <= 4-person rooms
> - 27 self-paid attendees in the 8-person rooms, only filled to 6
> - 9 sponsored attendees in the 8-person rooms, only filled to 6
Only filling these to 6 seems wrong to me. These are all more modern
rooms than some of the smaller rooms. So I would consider them
equivalent in comfort to the 5 and 6 person rooms even if filled with 8
persons.
> - 16 self-paid attendees in the medium/large sleeping bag rooms
> - 49 sponsored attendees in the medium/large sleeping bag rooms
>
> FWIW, if the numbers were different and showed there was more demand than
> availability for the small capacity rooms, I would support allowing
> sponsored attendees to pay a supplement for an upgrade. But unless the
> numbers change a lot more than I would expect over the next couple of
> months, I think the only fair way to do this is:
>
> 1) allocate self-payers to the class of room they've paid for
> 2) allocate the "non-communal" sponsored attendees to the small rooms
> 3) distribute the rest of the sponsored attendees evenly between all the
> rooms
>
I don't like this way of giving preference to those that said "No" to
the communal accomodation. This effectively changes the question to
"Would you prefer non-communal accomodation if offered for free?". I
guess everyone would have answered "Yes" on this. So you change the
rules after the game and give an advantage to some that answered "No" to
a different question.
I currently don't see a better solution to avoid this than selling the
upgrades at a moderate price. While selling will introduce the problem
that not everyone has the same amount of money to spend, it at least
makes the attendees back their preference with a payment and gives
DebConf some money we can use for the conference. If anyone has a better
proposal than selling the upgrades I'm listening.
I'm also open to consider free upgrades to 4 person rooms if there are
reasons why someone is unable accept communal accomodation.
> In particular, I think it is *not* reasonable to leave small rooms empty
> while sponsored attendees are crammed into the sleeping bag rooms simply
> because the budget calls for bringing more income from sponsored
> attendees.
I agree to this. I'm not sure if many small rooms will stay empty if we
offer upgrades. If so we could do a lottery for the remaining free beds.
> [1] I think someone moved the line for "communal" vs. "non-communal" after
> the earlier discussions, I consider 8 people in a room "communal" and am
> considering it such for this analysis. I'm pretty sure the distinction is
> purely semantic for purposes of this analysis; but if someone wants to
> re-run these numbers with "8 people per room" excluded, feel free.
I think we all agreed on the phrasing of this at the IRC meeting were
you, I and cate fixed all the details about the penta registration
options. I don't think this was changed later.
As a final note I also want to say that I think finding a solution
acceptable to the team soon is the most important thing. So I'm quite
open to compromises and won't insist on my position too much. So if a
majority thinks that vorlons proposal is fine, I'm fine with that too.
The worst outcome IMO is to delay the decision for more than a few days.
Gaudenz
[1]
http://meetbot.debian.net/debconf-team/2013/debconf-team.2013-04-08-18.02.log.html#l-93
(first discussed at 18:16, then "agreed" around 18:50)
--
Ever tried. Ever failed. No matter.
Try again. Fail again. Fail better.
~ Samuel Beckett ~
Reply to: