November 10, 2003

NFJS day 2

After being so timely with the day 1 summary I’ll have to play catchup for the No Fluff Just Stuff days 2 and 3 before the information seeps out of my brain.

And before you ask: any criticisms and praise I have in here (and in Days 1 and 3) have already gone into the speaker evaluations. Feedback is not just good for agile development, and if the speakers are going to change their presenations as a result of comments then we're helping our fellow developers down the line, right?.

Day 2 started off with Dave Thomas discussing mock objects. Excellent presentation that outlined not only what mock objects are, but also when and when not to use them. I was happy to see that he liked EasyMock for dynamic mocks, and that at least one of my co-workers attended. (I'm trying to get them infected...)

Next was two senior engineers from Sun (John Crupi and Dan Malks) discussing patterns. (Both are also authors of the recently updated Core J2EE Patterns book. They were fairly comfortable speakers (John moreso than Dan) and had a lot of experience to draw on as they went through the patterns they added to the 2nd edition of the book and how they connected to the rest of them. The disparity between the room and screen size made their graphics very difficult to read, and I would have liked to see more implementation discussion at the expense of covering all the patterns.

After lunch a panel with a number of the speakers answered questions from the audience. It eventually got to the "Can we trust Sun?" and "Will Sun be around?" area, which sucked for the one Sun guy (John Crupi) on the panel since he got pounded from the panel and the audience. Other than that there was entertainment all around, including the unintentionally funny discussion of user groups that Mark already mentioned. (Yes, we can be that juvenille...)

My third session was Stuart Halloway's 'Exploring Tests'. Basically it's learning about third-party libraries by writing code. Many people do this, including me. The interesting part is treating this code like all code you write: be able to automatically build and run it, keep it in version control, continuously integrate it. The point is that this code represents what you know about a library when this library gets updated you'd like to know what knowledge of yours has broken. Good stuff, and he is such a good speaker. (He has a paper online about it.) Two more co-workers attended, the testing virus spreads...

The last session of the day was about JDO. It was interesting, but not very much new besides what will happen in version 2.0, and nothing that will sway me into using JDO right now over Hibernate.

Saturday night we went out to a nearby Japanese place and had way, way too much food. Something about an expense account... After that we just came back to the hotel and watched a surprising Pitt victory over Virginia Tech.

Next: NFJS day 3
Previous: Managing packages for OS X