[series of entries regarding my trip to JIA]

Keith Donald's introduction to Spring was a presentation I had hoped would shed some light on the one framework that I'm truly interested in aside from RIFE and maybe JSF. Unfortunately this presentation had some good initial information but mired by, in my opinion, uninteresting slides and a problem with showing the entire ballroom Spring bean declaration XML for 30+ mins.

Spring certainly sounds interesting and I'm going to give it another look later this week at Matt Raible's Comparing Java Web Frameworks: JSF, Struts, Spring, Tapestry, and WebWork.

The thing that could've really perked this up was to show example code, and then show what it does in a browser. Back and forth. Coding... Browser... Really highlight the features, the best parts. Not trying to be unbiased and showing everything in just a factual way. If you're giving a presentation the audience expects that you have a fervent love or at least a mild passion about what you're presenting. I'm certainly not saying that Keith Donald himself doesn't have any interest in Spring, just that his presentation style suggested that.

Tragically, I was so uninterested in the style of presentation that I didn't actually learn what I wanted from the presentation. I've seen (and heard of) a few posts on JavaBlogs about the presentation and they had good things to say, perhaps it just wasn't made for my user group.

[series of entries regarding my trip to JIA]

Cédric's presentation was one I was looking forward to. As a user of TestNG, I wanted to put a face the the mailing list replies and to learn a bit more about a testing framework that I think has a lot of potential to go places.

Cédric's presentation was slightly different than the others so far. It seemed more intimate. It had a lot more geek humor. He slides were aesthetically pleasing with images from Battlestar Gallactica and SpiderMan. Which gave the audience a bit of a laugh each change and kept the momentum and the interest level up. Some of the images distracted a bit from the slides as they were large, but this wasn't as much of an issue as one would think.

The presentation started with an in-depth overview or a state-of-the-union of the testing world. He went over JUnit, both the current and the new not-yet-released JUnit 4, and a few other minor frameworks. This was news to me as I had not yet heard of JUnit 4 and it's new features. Cédric really did a great job of being un-biased when it came to being critical of JUnit. He then went on to address the problems that he perceived with JUnit and how TestNG addresses said problems. He also had some slides that showed some code and asked what we, the audience, thought would happen with each one and then why he thought it was a problem with JUnit. Everything was really well explained and thought out. Instead of telling us, like some presenters do, “this is the way it is” he presented the information and let us make up our minds.

He went through a few examples of use and then took a good deal of his remaining time to field questions. Of which some really interesting theoretical concepts came up, with some people challenging his assessment of JUnit and his implementation of TestNG. Cédric really stood by his decisions and the decisions of the other contributors.

I would've liked to have seen some more advanced examples and some useful TestNG patterns that had been talked about on the mailing list to be included. But all-in-all I was impressed with the quality of both the speaker and the content.

[series of entries regarding my trip to JIA]

Ramnivas really stepped up the level of the sessions with this one. I was extremely impressed with his delivery and with his content. He had a lot of concrete examples and he also presented common misconceptions that he explained. That seems to be the magic formula, concrete examples and myths. Those two things really make a session a lot more interesting.

Going into the presentation I had the assumption that while AOP was extremely powerful, it complicated things. By that I mean, you couldn't see everything that was happening in each method. What Ramnivas showed to countermand that myth was various IDE plug-ins that allowed you to see where your aspects would be applied in real-time. This myth really was one of the only things stopping me from using AOP, and seeing this made me download AspectJ for Eclipse that moment.

Another myth that was shattered was the idea that Aspects make your code harder to test, but Ramnivas makes an excellent point. If you use aspects, you can test the aspect, and then test the un-instrumented method and you've tested the entire thing. You get better separation and therefore have an easier time getting close to 100% on your test coverage.

One thing I've had a problem with so far has been the projector in the main ballroom. It seems that it's always blurry and if you don't sit in the first few rows you can't see anything at all. Some of the presenters have tried to alleviate this problem by having really gigantic fonts, with characters bigger than a man. But it only exacerbates the problem because they end up having to use the horizontal scroll to let you see all of the code.

But back to the AOP presentation, I really must commend Ramnivas on an interesting and engaging presentation that really sold me on the technology. I cannot wait to find an excuse to use AOP in my work now.