Pycon 2011

PyCon 2011
Conference Details
Atlanta, GA
2010 attendance: ~1100 -> 2011 attendance: ~1500
Many sponsors

PyCon 2011
Python Universe
Web Development
Scientific Computing
Networking
System Administration
Rapid Prototyping
Systems Testing
This talk focuses on the web development and sciencey aspects of Python

PyCon 2011
Why Use Python?
Elegant and concise syntax
Great documentation
Large standard lib
Tons of high quality 3rd party libraries
Useful collection of built-in data types
Cool features:
lambdas, list comprehensions, generators, properties, decorators
Avoids pitfalls of other dynamic languages:
Namespaced
Everything is an object
Strong typing with no implicit or explicit casting
Runtime error on undefined variables
Error handling is consistent: runtime errors are always exceptions

PyCon 2011
Python 2 vs 3
Python 2.7 is last version in 2 branch (for now…)
Python 3.2 was released in February
Many 1st level dependencies have been ported to 3
Expect more rapid adoption as the number of 3rd party packages grows
PSF is providing funding for open source projects to port
2 branch is still standard for application development

PyCon 2011
Interpreter (VM) Status
CPython: Reference implementation
2.7 – 3.2
PyPy (the new hotness): Pure python implementation
Uses a JIT to get better performance than CPython
Supports 2.7
Jython: Python on JVM
Supports 2.5
IronPython: Python on CLR
Supports 2.7
Possibly losing support for Microsoft

PyCon 2011
Web Development
Django is becoming the de-facto standard for Python web development
Huge community
Reusable application modules
Runs on GAE
PostgreSQL, SQLite, and MySQL still popular database options
MongoDB and CouchDB are popular NoSQL options
Memcache or Varnish for caching
Apache with mod_wsgi for server
JSON for AJAX
Pip and virtualenv used for dependency management
Celery plugin used for executing background tasks

PyCon 2011
Scientific Computing
Biology, physics, chemistry
Large scale parallelized simulations
90% Python, computationally intensive parts in C or Fortran
File handling, process management, networking, gui, data visualization, testing
Rapid prototyping of scientific algorithms and statistical computations

PyCon 2011
Parallel Computing
Celery: distributed job queue
Write tasks in Python
Tasks are executed on local or remote slaves
Handle results asynchronously
Integrates with Django
mpi4py: Python interface for MPI
Used at Argonne National Laboratory on 100K node machine
MPI is still around
mrjob: Python interface to Hadoop
Write map/reduce functions in Python
Submit jobs to Amazon Elastic MapReduce

PyCon 2011
Sessions
Session videos online at blip.tv (hopefully a less crappy site soon…)
Session descriptions at PyCon site
Sessions I attended:
Distributed Tasks with Celery
Javascript for people who know Python
Supporting All Versions of Python All the Time with Tox
Linguistics of Twitter
Mrjob: Distributed Computing for Everyone
Extreme Network Programming with Python and Linux
Rapid Python used on Big Data to Discover Human Genetic Variation
Python for High Performance Computing
The Data Structures of Python
Documentation Driven Development
Genetic Programming in Python
Exhibition of Atrocity
API Design: Lessons Learned
What would you do with an ast?
Through the Side Channel: Timing and Implementation Attacks in Python
An outsider’s look at co-routines
Best Practices for Impossible Deadlines

Pycon 2011

by limscoder

Accessibility

Categories

Upload Details

Usage Rights

2 Embeds 225

Statistics

Pycon 2011 Presentation Transcript

http://www.limscoder.com	222
http://translate.googleusercontent.com	3