Curious Venn

The last two days of PyCon AU were devoted to sprints.

There were active sprints for CPython, Django, Zookeepr, Requests, Persona, and various other projects. There were about seventy people on the Monday, and forty on the Tuesday.

Chris N described the Wrest Point Casino as a “Frankenstein” building, which is a good description of how different parts of the casino have been tacked onto the original 1930’s hotel over the last six decades. The conference was based in the old hotel, which has some beautiful art deco rooms. The sprints happened in the Derwent Room, which happened to have a very nice fireplace.

Cunningly, there were no power points near the fire place, so you coded there for as long as your battery allowed…

PyCon AU 2012 was held this last weekend (Aug 18 – 19) in beautiful Hobart. Here are some of my highlights…

There were two great keynotes by Mark Ramm and Kenneth Reitz.

Mark Ramm’s keynote

The biggest problem of PyCon AU? Too many good talks on at the same time! Apologies to the various friends whose talks I didn’t get to. Maybe next year? That combined with running three hours of tutorials across the two days meant I didn’t end up seeing as many talks as I wanted, but I did get to the following talks:

• Python Dark Corners Revisited by Peter Lovett
• Building a Cloud Service with Python by John Barham
• Python Lifesavers by Duncan Macneil
• PyPy – is it ready for production? by Mark Rees
• First time speakers lightning talks by Brianna Laugher
• Working with large scientific data at the Australian Synchrotron and beyond by Steve Androulakis
• Detecting and visualising large scale colour trends by Dennis Hotson and Lars Yencken
• Big data with python by Alex Sharp
• A noob looks at Message Queueing by Richard Jones

Quote of the conference?

>>> True = False
>>> True == False
True

Peter Lovett visiting Python’s dark corners

As always, I walked away with a long list of new tools, libraries and web sites to check out: MyTardis, Python Guide, Sikuli, R-Tree and a bunch of web tools: fabric, diazo, gunicorn, nginx, daemontools, celery, memcached, stathat.

There was a fantastic collection of inspiring, interesting and amusing lightning talks at the end of each, including talks on dancing robots, automating logins to banking web sites, Sikuli (again) and Python 3.3 highlights.

Another highlight was the cruise down the Derwent River to Peppermint Bay to the (yummy) conference dinner.

Derwent River cruise

Conference dinner menu

There were lots of other highlights: the conference was very well run, good food, lots of interesting chats in the corridors and between talks, and the early morning runs with conference people.

The view from the conference at dusk

And finally: excellent “African Swallow” coffee from Ritual Coffee in Launceston — thanks for the top-notch strong flat whites and the business card: I never realised that “Bean Shepherd” was a profession!

It’s now on to two days of sprints.

PyCon AU started this year with CodeWars on the Friday night.

The elimination round saw teams starting with a link to an EXE file that led to a puzzle of missing prime numbers. The first four teams with the correct answer went on to the semi-finals.

CodeWars warm-up

Recovering messages from PPM images, finding secret phrases, helping Olympic athletes take their drugs… it was a varied night of coding challenges.

CodeWars semi-final 1

CodeWars semi-final 2

The final code-down was between the Free Software Melbourne team and the “My Little Pony” team (I think that was the team name… it was late, there was wine, but I swear they called themselves “My Little Pony”).

The Free Software Melbourne team in action

My Little Pony was victorious in a battle of wits, wit, coding and an amusing Django app that almost worked.

These are my notes from the August 2012 Melbourne Python Users Group meeting at Inspire 9.

Andrew Walker (DSTO), Daniel Cousens (Swinburne) – “A grab bag of Pythonic computational geometry code”

Andrew and Daniel gave a tag-team test run of their talk for PyCon AU 2012.

Lots of meaty computational goodness, with plenty of interesting examples, including
numpy,
CGAL,
MayaVi,
and scipy.

On my (ever growing) to-do list is to check out
scipy.weave, which allows you to access C++ code within Python.

Other discussions…

Richard gave an overview of the PyCon AU conference this month.

Loki from Gravity4 discussed Fruit vs Robot which is due for release soon, and the joys of supporting hundreds of thousands of network connections with gevent.

These are my notes from the July 2012 Melbourne Python Users Group meeting at Inspire 9.

Bianca Gibson – Blender games using Python

Bianca demoed Whirlstrom, a game written in 48 hours, as part of the Global Game Jam competition.
The game uses Python scripting within Blender, with the logic and workflow defined by “bricks“.

Javier Candeira – OAuth: why, what & how (authorization and authentication with oauth and duct tape)

Javier gave an overview of OAuth, which is for robots; it’s like OpenID, but for authorisation not just authentication; summarised in the OAuth authentication flow diagram.

Javier gave a live demo of the rauth library.

Chai Ang – Getting started with Jython

Jython is at version 2.7alpha2 (30 May), but the stable version is still v2.5 and there was no v2.6.

It’s a good glue language for Java library code, and Chai showed examples from JDBC and Swing.

Jython is used by BEA Weblogic, IBM Websphere and more…

Jython is really only useful when you need to use Java libraries; eg using the (bettter?) Java XML libraries or deploying django inside JBoxss (for ease of distribution).

Graeme Cross – Command line argument processing showdown: a battle between four different ways in the standard library and also some PyPI modules

I gave a talk based on some of the prep work for one of my PyCon AU 2012 talks. If I pull my finger out, I might actually get around to posting the talk here! 🙂

Also:

• Julython is on!
• Barcamp Melbourne is coming up at the end of July

These are my notes from the Melbourne Python Users Group meeting this month at Inspire 9. Any errors are 99% likely to be mine in the transcribing or the understanding, rather than the speaker’s.

Tennessee Leeuwenburg – Udacity and Coursera: online courses

• Online tertiary-level courses, spun out of places like Stanford,  subjects cover computer science, science, statistics, etc
• Udacity
  • Courses are 7 weeks long, ~ 10 courses on offer, less than Coursera
  • Perhaps “friendlier” than Coursera, more “cool”?
  • Tennessee has done the AI course; loved it
• Coursera
  • Stanford, UPenn, Princeton, etc
  • Getting 100% “is the new pass”
  • Has a wider range of non-CS courses
• Courses start on designated dates, rather than on-demand
• 60 – 90 minutes of video per week? and 1 – 5 hours of work per week

Ed Schofield – What’s new in Python 3.3

• A straw poll of the room showed that, out of well over 20 people, no-one has done serious work with Python 3!
• Python 3.3 is in alpha at moment, aiming for a 25 Aug 2012 release date
• Some key changes:
  • Unicode: u'Hello' is back: PEP 414, helps with porting from Python 2
  • Space efficient Unicode: PEP 393
  • New packaging module: deprecates distutils, replaces pip, distutil2, etc
  • venv module: virtualenv-like support
  • lzma – new, compression algorithm
  • decimal module: faster, rewritten in C
  • bz2: rewritten, more flexible
  • mock: Mock objects for testing; now in the standard library
  • os module: lots of POSIX fixes
  • OSError: PEP 3151
  • Finer granularity in exceptions, with lots of subclassed exceptions
• Read the “what’s new in 3.3” release notes
• Significant libraries and their Python 3 compatibility status:
  • Django 1.5 is due in spring 2012 with experimental Python 3.3 support
  • Numpy supports Python 3 (has since v.1.5.0, Aug 2010)
  • matplotlib-py3k branch on GitHub, ported late 2011
• Plenty of hold-outs: boto, Fabric, South, gunicorn, Werkzeug, oauthlib, Flask & Jinja…
• The Who’s on Python 3? website tracks the top 50 Python projects and their Python 3 compatibility
• “6” – compatibility library for Python 2 and Python 3

Rhyd Olwin – Python in a hostile environment

• 550 users, but outside of IT, there are only 4 developers & no dev culture or environment
• How do you develop in a Windows shop with locked down PCs and no admin privileges?
• No exe’s can be downloaded/installed, but can use setup.py to download source code and build

Richard Jones – LightTable

• Different paradigm for IDE/coding
• On Kickstarter, has funded Python support as well as Clojure and SQL
• Have a look halfway through the promo video at the Flask demo

The next meeting is July 2 at 6pm.