I've attempted to summarize the discussion that occured yesterday on irc
beginning at 14:51 UTC. In case anyone is missing logs from that time,
see the attachment. -edrz
---------------------
edrz was adding timeline items from localteam meeting to
http://debconf10.debconf.org/dates.xhtml
moray was reviewing and asked: "Do we really expect to make sponsorship
decisions between the end of Thursday 15 April and some time on Saturday 17
April? seems a short window"
Ganneff and moray provided a description of the previously used process:
a "first round" of sponsorship decision normally requires ranking who has
requested it then giving out a first batch of money, it's normally been 95% of
the time taken
a geographically diverse team of 10-12 DDs was recruited. That team chose the
date/time for the meeting via doodle poll. coordination happened via mail &
irc. The meeting occured on irc.
Prior to the meeting, team members spent some time doing combined or
individual research on those requesting sponsorship. A csv of data is generated
containing: travel amount, travel data, name, mail, dates, person type.
One point of this research is that many applying will not be known beforehand
to the team, and it is good to avoid "never heard of them"-type rejections.
During the irc meeting the data can be posted, one potential sponsoree at a
time, each team member votes: -1, -0.5, 0, +0.5, +1 for example (or perhaps
larger range). after all requesters have been voted upon, votes are tallied to
create a ranking.
for those who end up with the same ranking, a second round of votes can be
held.
Rationale/considerations in favor of above:
avoid cabalistic decision process
ensure consideration of global differences
Voting criteria:
sponsoree involvement in debconf, debian, etc.
participation in DC: speakers, volunteers and members of the orga team
get some preference.
amount requested: if clearly outrageous sums requested, that counts negatively
in some cases, part of the total requested was offered.
Global team is not opposed to changes in the process, merely providing
information. In any case, a process and team need to be set RSN. Perhaps we
can set some limited time frame for alternative procedures to be proposed and
discussed. Then make a decision about how to procede.
Concerns voiced wrt to the past process/things still unclear/suggestions for
improvement:
Concern was raised about this being a potential shouting match or popularity
contest.
Clint stated the above process is not what he had imagined. It would be good
to hear from him how he envisioned it.
whether/how to consider the amount requested.
timing and dependencies between talk submission, sponsorship and reconfirmation
might be possible to keep some travel funds in reserve for speakers
Whether to pay all of a smaller batch vs. a % of a larger batch
differences in perception: some see DebConf sponsorship as a reward for all
their hard work on Debian. Others might not request funds because they _could_
pay even if it is hard for them to do so, but end up not getting any funding
because they are reluctant to ask for it. Defining "need" is difficult and not
really possible to document.
how to weight the ranking/fund allocation for those who are coming from
further away and therefore have greater expenses.
how to fold past sponsorship into the voting criteria
During 17-3 globalteam irc meeting shultmc agreed to lead the dc10 sponsorship
team. He has worked on this for previous years.
Data from munin posted by Hydroxide:
http://munin.debconf.org/debconf.org/skinner.debconf.org.html#Pentabarf
Currently the requested sponsorship so far totals just under $20600
we already have 12 people who have chosen to pay the $650 Professional fee, and
31 people who are attending without asking us to pay for food or accommodation,
and another 23 people who are not asking us to pay for one or the other. only
32 attendees are not requesting to stay in our lodging, and only 2 are
admitting they don't want to eat with us. the 23 are listed as only wanting us
to pay for either food or accommodation, but not both. (11 for one and 12 for
the other) and the 31 are not asking us to pay for either
198 registrants total
Moray:
for dc7 we had 283 'arrange own', '168 hostel sponsored' according to munin,
though I think it lacks some of the final detail whereas for this year we have
198 total, 156 'regular room' according to munin -- but the dc7
non-accommodated numbers went up a lot later on the main reason to register at
this stage is if you want the free stuff
Attachment:
full.tar.bz2
Description: Binary data