|
Mar 16
2008
|
PyCon 2008: Day 3 SummaryPosted by Mike Riley in Python, Programming Languages   |
 |
My final post for the PyCon 2008 conference coverage.
Aza Raskin gave the first keynote of the day, once again on the theme of 'Why Python Sucks'
Problems with Python include:
The main keynote of the morning was given by Mark Hammond of the Mozilla Project. His presentation, entitled "Snake Charming the Dragon: the past, present and future of Python and Mozilla", had several key takeaways, particularly regarding his experience with making Python work within Mozilla. He also talked about the future of Mozilla 2.0, focusing on the JavaScript 2.0/JIT compilation, the Tamarin ActionScript-based ECMAScript virtual machine. If Mozilla runs the xpcom library under Tamarin, much of the Python library will not be available. "Python continues to be used by, and influence the future of Mozilla. All it needs are some good friends."
Ivan Krisic presented an OLPC status update. The OLPC, or One Laptop Per Child, desktop shell is written in Python and the shell can be paused at any time to review the Python code running it. In the presentation, Ivan talked about how he has helped deploy the OLPC in Uraguay and Peru. Ivan recounted a story how he almost died while dismantling a satellite dish on a weak roof. Like walking on thin ice, the roof tiles began to crumble and gave way under his foot. Fortunately, the drop stopped at his ankle... but had it not, he wouldn't be speaking at the conference today.
Ian Bicking discussed how Python code can consume and parse HTML using the "BeautifulSoup" (optimal for screen-scraping), "html5lib" and the standard but awkward "HTMLParser" libraries. 93% of published web pages in the world are invalid HTML: tag soup, write-ony and presentational (not semantic). However, "HTML is the most important markup language in the world." XHTML still hasn't caught on after 7 years of effort. In Ian's opinion, XHTML only works well in 'walled gardens'. In Ian's opinion, lxml (a wrapper around libxml2) is the best of the various Python-friendly parsing libraries due to its good DOM support and its comparably fast performance. lxml.html, a class that Ian's been working on, extends lxml with an ElementTree model that provides simple representation for any tree, a diff function to identify page source differences, XPATH support
Next, Stephan Deibel demonstrated the commercially available Wing IDE for Python development. This product (writting in 84% Python, developed using Wing itself), available for the Windows, OSX and Linux platforms, looks really interested (especially for Django, TurboGears and wxPython development) and an IDE that I will be taking a closer look in the future, so stay tuned.
The last formal PyCon 2008 talk I attended was Mark Ramm's "Pylons and TurboGears: Working together on the web". TurboGears is a Python-based web application framework that competes again Zope/Plone and Django. While well attended, Django presentations and reception seemed to own this year's conference. Consequently, the momentum behind Django application development should really help propel this platform in 2008 in my opinion.
Lightning talks on a variety of topics were presented in the afternoon along, followed by a tutorial on sprinting for those interested in participating in the following week-long sprinting development sessions.
With this post being my final PyCon 2008 coverage write-up, several attendees and speakers alike commented to me that this year's PyCon was definitely the most active and engaging one yet. Nearly everyone I spoke with voted to attend next year's PyCon, also being held in Chicago. Perhaps by that time, Python 3 will have reached it final release milestone, Django will have attained its 1.0 release and the numerous projects that are underway will give future PyCon presenters and keynote speakers plenty of material to talk about and share in the future.
Aza Raskin gave the first keynote of the day, once again on the theme of 'Why Python Sucks'
Problems with Python include:
- Large download for Python base.
- Not easy to distribute Python scripts.
- Silo'd
- Good on Linux and OSX as the base is bundled with the distribution/OS, grim on Windows.
The main keynote of the morning was given by Mark Hammond of the Mozilla Project. His presentation, entitled "Snake Charming the Dragon: the past, present and future of Python and Mozilla", had several key takeaways, particularly regarding his experience with making Python work within Mozilla. He also talked about the future of Mozilla 2.0, focusing on the JavaScript 2.0/JIT compilation, the Tamarin ActionScript-based ECMAScript virtual machine. If Mozilla runs the xpcom library under Tamarin, much of the Python library will not be available. "Python continues to be used by, and influence the future of Mozilla. All it needs are some good friends."
Ivan Krisic presented an OLPC status update. The OLPC, or One Laptop Per Child, desktop shell is written in Python and the shell can be paused at any time to review the Python code running it. In the presentation, Ivan talked about how he has helped deploy the OLPC in Uraguay and Peru. Ivan recounted a story how he almost died while dismantling a satellite dish on a weak roof. Like walking on thin ice, the roof tiles began to crumble and gave way under his foot. Fortunately, the drop stopped at his ankle... but had it not, he wouldn't be speaking at the conference today.
Ian Bicking discussed how Python code can consume and parse HTML using the "BeautifulSoup" (optimal for screen-scraping), "html5lib" and the standard but awkward "HTMLParser" libraries. 93% of published web pages in the world are invalid HTML: tag soup, write-ony and presentational (not semantic). However, "HTML is the most important markup language in the world." XHTML still hasn't caught on after 7 years of effort. In Ian's opinion, XHTML only works well in 'walled gardens'. In Ian's opinion, lxml (a wrapper around libxml2) is the best of the various Python-friendly parsing libraries due to its good DOM support and its comparably fast performance. lxml.html, a class that Ian's been working on, extends lxml with an ElementTree model that provides simple representation for any tree, a diff function to identify page source differences, XPATH support
Next, Stephan Deibel demonstrated the commercially available Wing IDE for Python development. This product (writting in 84% Python, developed using Wing itself), available for the Windows, OSX and Linux platforms, looks really interested (especially for Django, TurboGears and wxPython development) and an IDE that I will be taking a closer look in the future, so stay tuned.
The last formal PyCon 2008 talk I attended was Mark Ramm's "Pylons and TurboGears: Working together on the web". TurboGears is a Python-based web application framework that competes again Zope/Plone and Django. While well attended, Django presentations and reception seemed to own this year's conference. Consequently, the momentum behind Django application development should really help propel this platform in 2008 in my opinion.
Lightning talks on a variety of topics were presented in the afternoon along, followed by a tutorial on sprinting for those interested in participating in the following week-long sprinting development sessions.
With this post being my final PyCon 2008 coverage write-up, several attendees and speakers alike commented to me that this year's PyCon was definitely the most active and engaging one yet. Nearly everyone I spoke with voted to attend next year's PyCon, also being held in Chicago. Perhaps by that time, Python 3 will have reached it final release milestone, Django will have attained its 1.0 release and the numerous projects that are underway will give future PyCon presenters and keynote speakers plenty of material to talk about and share in the future.
Subscribe to this comment's feed
Show/Hide comments
Show/Hide comment form
