[Conferences] PyCon Argentina 2009, finished!

[Facundo Batista]

(disclaimer: there's a Spanish version of this mail in my blog [0]...
If you read Spanish I encourage you to go there because not only it's
better written, but also images and links are as they should be, not
footnotes)

Ok, after so much foreplay and expectations, we finally had PyCon
Argentina 2009, the first Spanish-spoken Python National Conference.

My idea here is not comment how the event was from an end user point
of view, but try to transmit feelings, thoughts and ideas about the
organizing/executing side.

(image: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-cierre.jpg)


Event Format
------------

The event was during two days, morning and afternoon, with three
tracks in parallel most of the time. We used three auditoriums with
different capacities (the biggest with 300 seats).  Presentations were
of the following types:

Talks: They happened in the three auditoriums, with national speakers
mostly, always in Spanish. Also there were some cases with most of one
speaker, like a panel. The assigned time was 40 minutes, plus 5 of
Q&A. After each talk, we left 15 minutes for people to go to the
bathroom or another auditorium.

Lightning Talks: These are 5-minutes talks. Anybody can give one of
these (people writes their names in a paper during both days), with no
restrictions in the subjects. Normally you don't expect too much
quality of these talks (because they're fast, and tend to be
improvised), but having 10-12 different subjects in only one hour
makes it valuable. For these talks, all people seat in the main
auditorium.

Plenaries: We had two plenary talks (in the main auditorium, with all
the assistant public, the whole hour), that were given by two
international invitees: Jacob Kaplan-Moss (creator of Django [1], one
of the web frameworks most widely used in the world), and Collin
Winter (Unladen Swallow [2] developer, a Google project that promises
a Python 5 times faster than the actual one).

(image: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-muchagente.jpg)


Team Organization
-----------------

Python Argentina does not have any formal or informal structure, we're
all peers and each one pushes the project that wants.

In this particular project, I structured the working team according to
a military hierarchic organization (the same used in the enterprises).
Basically I was the General Coordinator of the event, had a second
person of trust (one person that you fully trust, not only about which
decisions he/she makes, but also how he/she execute them; in this case
he was Alejandro Cura), a team of about 6-8 people, and a lot of more
eventual collaborators, with more or less participation (helping in
the conference days, or in different committees, etc.).

The most important requirement for a successful participation in the
organization, taking into account that we're all volunteers, is to
comply with what promised, no matter if the task is big or trivial. If
you say "I'll do it", *do it*.

The kind of work that is done throughout the year is very different
than the work done during the conference days. During the year the
response time is a lot lower, you need to take decisions that are
influent in the long term but can be discussed in several talks or
meetings, etc. During the conference days, the decisions are taken and
executed at the moment, and there's no time to doubt. Generally, the
better work done during the year, the conference days will be
smoother, with less surprises.

(image: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-azules.jpg)


Risks
-----

There are several important tasks that take a lot of time and bring
risks to the conference organization:

The place where the conference is done, and the services offered
there. In our case the conference was hosted in a private University,
to whom we payed for different concepts (internet, security, sound).
We've been talking with them since April about all that was needed,
but a lot of stuff happened too close to the conference date. In
particular, the internet access was installed the previous afternoon.
Luckily, we had the volunteer service of Buenos Aires Libre [3],
organization that set up the WiFi access, which was excellent.

Printing. Beyond the paper stuff that we printed or photocopied (ads,
polls, etc), we had three big things to print: the program of the
event (a 32 pages book, full color), the Python tutorials in Spanish
(116 pages book), and some Canonical notebooks. The design of each
thing takes a lot of time, and also takes time to print them. I saw in
too many conferences the mistake of sending everything to print too
close to the final date, and always there's something that is not
ready for the conference. It's always important to speak with the
printing people in advance, and involve them in the design, to
minimize later surprises.

Financial and economical aspects. Months before the conference we
created three different budgets, that we'd execute one way or the
other according to the funds raised later (and based on these budgets
we defined the sponsors categories). Luckily we didn't have to knock
any door to raise the money: all the sponsors were brought by people
that is in Python Argentina (because they work in (or own) the
companies that sponsored the event). There are two important points:
one is that the budget must be designed in levels, to maximize the
benefit for the conference if you can not raise all the money; the
other is that you need to start collecting the money as soon as
possible, because spending the money (say, execute the budget) is not
trivial and takes longer than you may think.

Technical infrastructure for the organization. There are several tools
that we used for us organizers to work together. We had a wiki to keep
everything annotated, a mail list as the main information exchange
mechanism, and IRC for some non-physical meetings that we had
(especially for the talk selection, where participated a lot of people
that live in other cities). We also used a SVN repository to keep all
the files, specially the web site files. For the web site of the event
we used pycon-tech [4], a project thought for this kind of events.

The team. This is one of the main risks, because if you don't get a
group with people that tries to achieve the same goal, that wants to
get together and work towards a common target, it may happen that the
event is not well organized just because of the organizers themselves.
It's important to understand that in a conference organized by
volunteers, it may be people that starts in the organization and then
quits, or that people may jump into the wagon later. And the team is
one of the first things you need to define! This is why is critical
that the General Coordinator of the event to have enough power in the
community to have a team assembled and working together during all the
months that take to prepare the event.

Generally, is recommended to do what you know how to do, and ask help
to (or hire) specialists in the cases where you don't have the needing
knowledge or resources.  For example, the event wifi was set up and
operated by Buenos Aires Libre, the same team that handled the
wireless network in Wikimania 2009.

After all said, however, you may find that there are problems you can
not prevent or mitigate: the first day of the conference the
electricity went out in all the neighbor. Some talks continued in a
dark room for a while, and then we all went to have lunch, and lights
came two minutes before the afternoon session started...

(image: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-remera.jpg)


Content
-------

The content is determined by the plenaries and normal talks, because
the lightning talks don't go through the decision of the organizing
team.

For the international invitees plenaries, normally you choose the
invitee and give him/her total freedom about what to say. In our case,
the selection was done together with PyCon Brazil, and it was
amazingly well done.

On the other hand, for the normal talks, a Call for Talks is done some
months in advance, and after the presentation limit date, a committee
gets together to discuss which talks will be chosen. This selection is
based on the talk quality, and in the subjects that need to be
represented during the conference.

(imagen: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-collinwinter.jpg)


Social Aspects
--------------

One point that you need to take care and pay attention in the
conference is its social side. A National Conference is a very good
opportunity for people to get together, see each other again after
some time, exchange ideas and knowledge, etc.

The evenings after each day of conference is a good moment for people
to have dinner or some drinks together, and make this social encounter
happen.

In our case, we wanted to make this deeper by organizing an asado [5]
on the Sunday after the conference, to which all speakers and
organizers were invited, with their families. The goal here is to
enjoy a relaxed day together, eat good meat (or vegetables!), have a
glass of wine, talk, play some games, etc.

(imagen: http://www.taniquetil.com.ar/facundo/imgs/pyconar09-asado.jpg)


[0] http://www.taniquetil.com.ar/plog/post/1/425
[1] http://www.djangoproject.com/
[2] http://code.google.com/p/unladen-swallow/
[3] http://www.buenosaireslibre.org/
[4] https://pycon.coderanger.net/
[5] http://en.wikipedia.org/wiki/Asado

-- 

.    Facundo

Blog: http://www.taniquetil.com.ar/plog/
PyAr: http://www.python.org/ar/