Day 3 of the OSDC conference, I presented two sessions on Agile (with Paul King) and Data Migration, so preparations and organisation got in the way of attending and as many sessions as I would have liked. Here are my notes from the sessions I attended.

Keynote – Google Wave

Dhanji Prasanna presented the keynote, here are my notes from the session:

• answer to question “what would email look like if it was invented today?”
• not just about text, about structured content
• gadgets – state is stored in the wave (eg. using maps)
• can exchange waves across different wave servers using federation
• keep data you own on your own wave server, put you collaboration on the public wave server
• robot participants are first class participants
• spelling robot corrects spelling in context, and runs server side (Spelly), Linky does similar for links
• use open source tools like Guava libraries, Guice, WebDriver, GWT, OpenJDK
• all of Wave will be open sourced, including the client
• can play with the reference implementation right now

Experiences From Agile Projects Great & Small

The session I presented with Paul King had a good crowd in the main auditorium. The slides are available in a separate post.

“Change Bad!” – Change in Database Schemas and Source Code

Andy Todd delivered this session, here are my notes:

• “Data matures like wine, applications like fish”
• What changes – tables, views, stored procedures
• could put DDL in version control – not same as database object
• could put database independent model in version control – representation only
• database = DDL + data + implementation
• could store complete DDL for each scheme version – lots of work for DBA
• could store complete initial DDL & subsequent deltas
• could store complete DDL for every version and subsequent deltas
• don’t forget about reference data and schema (user) creation scripts and store in version control
• even if scripts are in version control, you should still have a backup
• secret to success is testing – should be like code and have unit tests and be automated
• hard to rollback
• every migration should maintain data integrity
• pioneering work by Scott Ambler – Database Refactoring
• high availability is a different kettle of fish – all of the above is invalid
• need to think about how you version control your database structures, different to how you manage source code

Data Migration In An Agile Open Source World

The session I presented had a decent crowd. The slides are available in a separate post.

Lightning Talks

Open Source Game in 5 Minutes

• using Rabbyt and Pyglet (Python based)

local::lib::environment

• working with multiple distinct libraries in Perl
• integrates with CPAN, etc…
• virtualenv in Python does the same type of thing
• cloning this for Perl, available on github

Ext.Direct

• no Perl implementation for ExtJS, so wrote Ext.Direct
• similar ports for a lot of different languages

OSIA (Open Source Industry Association)

• local industry association for open source
• $50 to join
• numbers used to give to government

bluehackers.org

• for geeks dealing with depression

Locknote – Open Source in Government

Pia Waugh delivered this keynote, my notes are as follows:

• Parliament House has 4,500 rooms (7,000 with bathrooms and kitchens!), about 8,000 people when parliament is sitting
• government is a very risk averse environment, a lot of fear
• Kung Fu Panda is a lot like parliament, not a lot of assistance from your colleagues
• dis-empowered community due to a lack of ICT skills
• Tech Support Cheat Sheet
• biggest digital divide is in accessibility and skills
• need to make ICT issues known to members of parliaments (facepalm)
• Government 2.0 taskforce – how to open up government
• data.australia.gov.au (also data.nsw.gov.au)
• Public Sphere