Once again, we rode all the way down to the PyConES Conference. This time we took a train to Almería to meet with the Python community. A bit closer to the sea. And just a bit warmer.
Good ideas and practices don’t belong to a specific language. What’s more, we think all communities should learn from each other. And that’s what we did. We found a good amount of talks that made a lot of sense to us.
Here’s a brief. Obviously, it doesn’t mean that the other talks were not interesting. It means either it didn’t click on us, or we could just not assist to the talk at all because we were in another one.
Some talks we found really interesting
Sistemas distribuidos en Python by Guillem Borrell
Focused on more advanced users, this workshop was a really intense and condense demo about making an scalable server using parallel threads and high performance tasks. Guillem is such an exceptional mind and we were overwhelmed from his experience.
Hola mundo, desde el fichero a la pantalla by Jesús Espino @jespinog
A deep insight about the compilation and execution processes that take place between your typing and the execution results. This one happened to be a very well explained talk about a not so common yet interesting topic, given by a great pythonista: how the compiler and the runtime work, covering concepts such as "the tokenizer", "syntax trees" and others related to compilers theory. A must-read Jupyter Notebook if you want to deep inside Python compiler and its runtime environment.
Heydjbot: Spotify & Telegram. Una historia de amor by Jose María Álvarez
Refreshing talk about the experience of Jose María building a bot that interacts with Spotify and Telegram messaging service. “A bot is a helpful piece of software that performs repetitive and daunting tasks a human doesn’t have to do”. And it made us think about making our own one!
Breve introducción a los algoritmos genéticos by Siro Moreno
Following the talk of our beloved Federico Mon (@gnufede) about actual machine learning: “Can a machine learn how to program?”, we were delighted with a deeper definition of what a genetic algoritm is. Many, many thanks to both. It was inspiring.
Efficient Django by David Arcos @DZPM
There are few talks about this topic with this level of detail and pragmatic approach. From database to web optimisation. It’s good to remember the old tricks and to learn some new ones too. If you missed this one, check the slides online: it will worth to take a look into them once in a while...
Python for developing a real time automated trading platform & osBrain, a general purpose multi-agent system module by Miguel Sánchez de León
These two talks together conformed a great:exclamation: introduction to osBrain, a general-purpose multi-agent system written in Python, and its main use case, a real-time automated trading platform. It was really interesting having two talks about such a different paradigm of "software agents". Although osBrain itself misses some key points from FIPA protocols and standards, it do is an interesting point to start with if you want to try the Agent-oriented Programming (AOP) while enjoying your skills in Python.
- "Python for developing a real time automated trading platform"
- "osBrain, a general purpose multi-agent system module"
Python y DTrace by Jesús Cea
Surely one of the most advanced talks we cared to attend this year. This talk by Jesús Cea taught was about instrumentalization and live debugging an already running python process using DTrace, a technology not yet available for all the mortals.
Cómo crear tu propio Shodan en casa by Jorge Coronado
It was surely an amusing and enjoyable talk. Jorge introduced us to a new tool —a Shodan like library— of his own that tracks unsecured websites, and analyses and stores them port by port for further investigation. It reminded us that security talks can be fun.
Los closures que emocionaron Spielberg by Víctor Terrón y Pablo Galindo
This was one of the most engaging talks we have seen this year, hands down. We started from a common yet weird bug that occurs when using closures in Python, and from there, we moved through the correct implementation. It was an advanced talk, so well performed and orchestrated that it looked like a fundamentals one. Closures are something we should all understand, but it’s always a pleasure to stop and get a ride through different cases, so we get to fully understand what is actually happening. Nobody is going to get a UnboundLocalError exception anymore. At least not without knowing why and what to do!
We found also interesting
- “Scaling Django: Running Django in Kubernetes”, by Ivan Pedrazas.
- “Channels is coming: real-time web in Django”, by José Manuel Rivas.
- “No lo abandones: él nunca lo haría” (lighting talk).
- “Depurando como un pro”, by Javi Romero.
- “Beautiful Python”, by Alicia Pérez.
- “MicroPython”, by Jesús Cea.
- “Cómo construir un API del que tus padres se sientan orgullosos”, by Adrián Matellanes.
Our talks at the conference
Fede is a passionate Python Developer who wanted to make his own research about machine learning, or rather "machine coding": '¿Puede la máquina aprender a programar?' (Can a machine learn how to code?). Following this sci-fiction topic, he brought us some conclusions of his own about how far can we get with genetic programming and neural networks. Using Python Abstract Syntax Trees as program DNA, he was able to randomly generate some valid code that passes predefined unit tests. However he ended up with a unexpected and controversial conclusion: they can't... yet. So it's not that straight forward to force automated software to work for us. They are not able to deliver as good results as we do using our complex mind, even if we try to replicate it for them.
It was an introduction —to many— to genetic programming. And for some of us, it was inspiring.
Thanks to the PyConES team for holding such a fun conference. We know this takes a lot of effort from you guys. All we can give you in return is our gratitude. We enjoyed it a lot. Really.
See you next year at PyConES 2017!