Oct 30, 2013
Ploneconf 2013 - Brasìl, an amazing experience of life
A detailed journey log of the latest Plone conference. A very interesting conference followed by two sprints unique for level of productiveness and awesomeness
If you have never attended a Plone conference, you are really missing something worthy.
There are always two main aspects in these events, the technical part and the social part.
Usually the technical part of the conference takes place during the day, but it's during the evening/night that the Plone community shows the "beauty behind itself". Those moments are when the community renews the links between members, when new friendship starts and when the fun begins. I decided to make a very detailed report of the latest conference in Brasil, in the hope that new members will join us as a new developers and, hopefully, as new friends.
The RedTurtle crew, Stefano and me, departed to join the conference place on September 30th. We started our journey from Bologna, and we arrived in London. We enjoyed the stay there, where everything works "britishly" smooth. After few hours, and after a really disappointing lunch, we took the connection to Miami. In 6 more hours (2 of them passed enjoying the "welcome" of the airport security), we finally took our flight to Brasìlia.
We arrived in Brasìlia in October 1st and once there, we met four other plonistas: Tom "Spanky" Kapanka, Cris Ewing, Clayton Parker, David Glick. After few greetings with the plonistas all togheter we took the bus connection to the hotels district.
The very first two things that we noticed there were: 1- no Brazilian seems to speak or even understand English, 2- the ground in Brasil is insanely red... yes, it's really red! Weird!
I don't speak Portuguese!
- the most useless sentence ever in Brasìl
The same evening, as a well established habit, we joined all the other plonistas that were already in Brasil at a pub. From that point until the end of the conference, we started to enjoy the two national specialties, meat and Caipiriña.
The day after was the first day of the conference. Stefano and me decided to take a walk to get to the conference venue. After some run for our lifes on the brazilian streets, and after being astonished for the hugeness of the new football stadium for the next world championship, we reached the place. Once there we got our conference pass and the fun began. There were 4 tracks of talks and usually 2 were in Portuguese (python-br) and 2 in English (Plone).
October 2nd - the fun begins
As usually, the conference started with a welcome keynote, at first from members of the national Python community and then from members from the Plone community. For the talks in Portuguese there was a quite good translation service too.
After a little break, the real fun started, with one of the most hilarious plonista out there, Mikko Ohtamaa. He talked about his love/hate for Plone and for the community.
After that I followed an interesting talk of a brazilian university (UFRGS) employee that showed us how they managed to build an infrastructure of more than 250 Plone sites and keep them organized in a single hostname. The 2 tools that they used were a bunch of nginx rules and the virtualhostmonster. I liked the talk and I think this should be done more often, in many public administrations in Europe too.
The last talk of the morning was hosted by Guido Stevens over "High-availability, high-performance" for Plone. During the talk he professionally explained how he managed to reach a 100% uptime, mission critical system with Plone. The main idea is to remove all the SPOFs (single points of failures) of the common Plone architecture and replace them with something else more resilient. The end architecture comprehends 2 twin blade servers with and hardware load balancer (or in alternative 2 haproxies), each of them with Apache+woodan (for persistent disk cache), varnish and haproxy (as a load balancer for the zope instances), N zope instances, and mysql with relstorage (instead of the zodb).
In the afternoon I decided to take a break from the talks and I went to the open space, where Franco Pellegrini kindly helped me getting started with Plone mockups and with a templating system developed by him, called "generator-mockups". This is a node package based on yeoman, that helps you getting started with mockups by creating a skeleton of a package (more or less like paster for Plone packages).
Then I followed the awesome talk of Sven Strack on the state of the art of the multiple ways for installing Plone.
In the end of the first conference's day, there was the Plone Foundation meeting were the previous board made a resumee of the last year of the foundation life and the new board was voted (well, actually it wasn't really voted 'cause the number of participants was exactly the number of chairs, so it passed automatically).
My second day at the conference started with the Matt Hamilton's talk about Plone's pluggable authentication system. He introduced the PAS in all its interesting aspects, he explained how to get started with our first pas plugin and then he presented three interesting pas plugins:
- pas.plugins.velruse: a PAS plugin developed by RedTurtle (hehe :) ) that lets you easily authenticate users from social networks
- netsight.windowsauthplugin: a PAS plugin that authenticates users against Kerberos systems (the most known example is active directory)
- netsight.aspxauthplugin: a PAS plugin that allows Plone to trust the auth of a .NET application and vice-versa (by encrypting/decrypting the authentication cookie).
After that, Paul Everitt (yes, the Zope's father) gave a . With his talk "How it happened", he did a chronological report of his working life, with mistakes and lessons learned, passing through the Zope birth up to his latest effort, Pyramid.
Another very interesting talk of the day was the one from Eric Brehault (the maintainer of Plomino). He basically presented his attempt to import Wikipedia into Plone without loosing much performances and using as tools Plomino and a new addon called souper. He also made a live presentation of the resulting import and it was quit impressive to see 5M of records into Plone that load in few seconds
October 4th - the conference ends
And there we came to the last conference day. In the morning Stefano gave his talk "Plone as a tool to promote citizen-driven public policies". He presented a case study with particular emphasis on the workflow of the whole project. He started from the analysis moment, went through the meetings and the methodoligies useful during every moment of the project. In particular the focus was also on the engagement of the stakeholders (the customers included).
In the afternoon Philip Bauer organized an open space to talk about the future of Plone in 5-10 years from now. The objective of the talk was clear, let's not talk of Plone 5 and all the cool stuffs that will be delivered soon, but let's find a long term vision. The common feeling was that everyone his concerned about the future of Zope because of its stale state and its inability to accept help from us. In the start of the discussion there were 3 options:
- we could do nothing
- we could bring Zope ahead and maybe fork it
- we could move Plone to another application server (like for example SubstanceD).
As the discussion took place it was clear that the first option was not viable. Between forking Zope and moving Plone to another appserver there was no clear winner, and truth be told this was far beyond the purpose of the discussion. The real purpose was to present the problem so that in the future there will be some discussion and we hopefully arrive in 5 years from now with a solution.
October 5th - the Sprint
The conference is finished but the fun is not. On Saturday we, the Plone community, moved to another venue and we started sprinting. That was one of the most productive sprint that I've ever attended. Here you can find the document used to organize each topics:
Here some of the most active topics of the sprint:
Timo Stollenwerk led this topic and presented his effort to benchmark Plone among its versions and with the diazo the applied/non applied. He also gave a presentation of a tool called jMeter that let you benchmark performances.
This topic too was very active. Mikko, Sven, Steve and me, we talked about the new layout that should have the documentation page in plone.org and we decided to move the collective.usermanual to a new subdomain: docs.plone.org. Looking at the great effort of Asko Soukka that used robotframework to autogenerate screenshots for some part of the documentation, we also decided to use it massively and to move all screenshots to that method. This will be a great thing because you, Plone distributor, will be able to build your documentation with your own Plone site running and using your custom theme. This obviously will permit you to give your customers a shimmering customized version of the documentation.
The answer to life the universe and everything is NIX!
- Rok Garbas
Steve McMahon proposed this topic to talk about how easy/hard is for a new developer to approach Plone and what we can do to ease this process. The discussion covered fields like how easy is to install Plone, how easy is to start development with tools like zopeskel/templer and mr.bob, and the future of plone.api.
Rok Garbas led this topic on the future of frontend development in Plone. He gave a bootstrap training on mockups in the hope to bring more developer into that. The training was very crowded and the rest of the sprint brought a high quantity of pull requests of bugfixing.
Stefano Marchetti led this topic in the effort of moving forward PloneGov. They decided to bring new life to the project.
Nejc Zupan was the promoter of this topic for bug fixing plone.api and to bring it to the very first stable release.
All good things come to an end
Time for us to get back home. We really missed the Caipirinha sprint but work called us. As always, joining the conferences of the Plone community at conferences is an amazing experience of life. See you all in few weeks in Arnhem Sprint. You guys rock!