I attended most of Brent Friedman’s Java and PostgreSQL talk. I do a little Java hacking for a legacy app, so the sysadmin-ish details about installation and upgrading were useful as I look forward to upgrading.
I really, really enjoyed Andrew Sullivan’s talk on Rapid Development as Denial of Service Attack. Not only is Andrew witty and sharp, he just nailed the people problems that are inherent in modern software design. I really liked his presentation style as well – tons of useful analogies. Some key bits:
* Be sure to define what “normal” is – so that you can tell when things are not normal
* Build in time for instrumenting – because you’ll either plan to do it up front, or spend unplanned time later
* Databases are really an API, not a black box
Greg Sabino Mullane provided a survey of all the replication options available to PostgreSQL, including the pros and cons. This sparked a 20-minute debate among some long time PostgreSQL contributors and a few newer consultants and conference attendees.
The discussion revolved around what replication technologies to include in the core of PostgreSQL. In particular, the support for statement-level replication was in dispute, as well as the utility of self-modifying code. Greg also mentioned that Bucardo’s configuration engine uses Moose, a Perl object framework.
To wrap up the day, I led a discussion of women in open source, along with Dr. Kathryn Bartol,
Yixin Qiu, Dr. Louiqa Rashid, and Laura Thompson of Mozilla Corporation. My slides are below. Kay and Yixin also presented preliminary results from their research into activities of the KDE Women group. They were studying communication patterns and developing labels for the different types and quality of communication occurring between people on the mailing list.
I look forward to their work being published. I’ve attached my slides below. I have a few links to add before I add the final version to conference site.
A few more photos are also available on the Flickr Pool.