I arrived to San Francisco yesterday and OOW kicked off today. The efficiency at the registration alone was impressive. The lines we’re very long, but it probably took less than 15 minutes to get through it.
This post will essentially be a dump of notes from the sessions I attended.
The first session was the start of the APEX symposium. It was supposed to feature Scott Spendolini and an overview of APEX 4.0. He had been delayed so the presentation was held by Tim st Hilaire and Doug Gault. It was a nice review of focusing on user interfaces. Technology is nice, but in the end it is all about how productive the end-user is.
They looked closely at the interface netflix has for managing the movies you want to see and used their own version of it for a troubleticket system. There was quite a bit of discussion about the problem with getting the business involved in prioritizing and how they built the solution to make it easy to get the busines side involved in making the right things prioritized.
The technical solution uses JQuery which is now part of APEX out of the box. They have placed their work in a plugin that can be downloaded from apexplugin.com.
The next session was SQL Developer Best practices. It went through the presenters favorite things with SQl Developer.
It covered a lot of things, the following are the ones I feel I need to use more.
Creating code templates and adding a hot key for activating it would eliminate rewtiting similar code over and over.
There is a way do generate documentation for a schema in HTM and other formats.
It includes a function for monitoring SQL and sessions that will reduce work with querying v$session and related tables to see what is going on in a system.
Unit testing is included in version 2.1. You create a repository and define tests that can be executed both from SQL Developer and from the commandline (ututil).
The third session of the day was “Messed up Applications” with Carly Millsap. It was an interesting session that wasn’t very technical for a Millsap presentation. He reviewed three systems that had been designed to perform poorly. Of course neither was designed that way on purpose, but they were all built in a way where tuning would not help.
I missed the first whoch was about something he had built himself. The next was a ticket reservation system that spent over a minue to bring back all possible trips a user could go on, when the user almost always chooses the first one. The interface presented him with 2136 trips to choose from. This can of course not be fixed without adjusting how the system interface works.
The last was a ramp anti-pattern for a company that had a system that ended up printing every invoice since day one. So day one it printed all invoices for day one, day two it printed all invoices for day ane and two and so forth. This of course wasted lots of paper and lots of resources very soon.
The last presentation was about trending performance and capacity through using AWR data. The presenter had reverse engineered the SQL in AWS and then used that to bring out data for different times for the same execution or the same SQL over time. It takes some work to do, but it increases the value of AWS significantly.
After that was the keynote, but that will be in its own post.